University of Southern California Dissertations and Theses

Some Effects Of Monetary And Fiscal Policy On The Distribution Of Wealth

(USC Thesis Other)

Some Effects Of Monetary And Fiscal Policy On The Distribution Of Wealth

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content SOME EFFECTS OF MONETARY AND FISCAL POLICY
ON THE DISTRIBUTION OF WEALTH
by
Elliot Allan Ponchick
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Economics)
February 1973
INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While
the most advanced technological means to photograph and reproduce this document
have been used, the quality is heavily dependent upon the quality of the original
submitted.
The following explanation of techniques is provided to help you understand
markings or patterns which may appear on this reproduction.
1. The sign or "target" for pages apparently lacking from the document
photographed is "Missing Page(s)". If it was possible to obtain the missing
page(s) or section, they are spliced into the film along with adjacent pages.
This may have necessitated cutting thru an image and duplicating adjacent
pages to insure you complete continuity.
2. When an image^on the film is obliterated with a large round black mark, it
is an indication that the photographer suspected that the copy may have
moved during exposure and thus cause a blurred image. You will find a
good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being
photographed the photographer followed a definite method in
"sectioning" the material. It is customary to begin photoing at die upper
left hand corner of a large sheet and to continue photoing from left to
right in equal sections with a small overlap. If necessary, sectioning is
continued again — beginning below the first row and continuing on until
complete.
4. The majority of users indicate that the textual content is of greatest value,
however, a somewhat higher quality reproduction could be made from
"photographs" if essential to the understanding of the dissertation. Silver
prints of "photographs" may be ordered at additional charge by writing
the Order Department, giving the catalog number, title, author and
specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as
received.
Xerox University Microfilms
300 North Zeeb Road
Ann Arbor, Michigan 43106
I
73-18,836
PONCHICK, Elliot Allan, 19H6-
SOME EFFECTS OF MONETARY AND FISCAL POLICY ON
THE DISTRIBUTION OF WEALTH.
University of Southern California, Ph.D., 1973
Economics, general
University Microfilms, A XEROX Company, Ann Arbor, Michigan V
• id
THIS DISSERTATION HAS BEEN MICROFLIMED EXACTLY AS RECEIVED
UNIVERSITY O F SOUTHERN CALIFORNIA
TH E GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES. CA LIFO RN IA 9 0 0 0 7
This dissertation, written by
...... ElliotMlan^P
under the direction of h.j,S„ Dissertation Com
mittee, and approved by all its members, has
been presented to and accepted by The Graduate
School, in partial fulfillment of requirements of
the degree of
D O C T O R OF P H IL O S O P H Y
Dean
Date.
February 1973
DISSERTATION COMMITTEE
/y^J ( f i i
Chairman
ACKNOWLEDGEMENTS
I wish to thank the members of my committee, Dr.
M. DePrano, Dr. G. Tintner, and Dr. R. Ward, for their
assistance in guiding me through this dissertation. I
am especially grateful to Dr. DePrano, the chairman, for
the many important suggestions he made, his inexhaustable
enthusiasm and his unfaltering patience throughout the
writing of the dissertation. I have benefited greatly
from his interest and friendship all through my graduate
program.
I am indebted to Ms. Evelyn Hansmire of the In
stitute of Survey Research of the University of Michigan
for her kind assistance in obtaining survey data for
this project and to the Survey Research Center for sup
plying this data.
I would like to extend my thanks to Mr. Jeff
Bahr for his generous and expert assistance that made
reasoning with the computer possible.
I am also grateful to my colleague Mr. Gopal
Kadekodi for helpful suggestions at each stage of the
dissertation.
Finally, I wish to acknowledge the assistance of
Mr. Helen O'Connell and Mrs. Pat Anderson who trans
ferred a set of notes into a finished manuscript.
ii
TABLE OF CONTENTS
ACKNOWLEDGMENTS ............................
LIST OF TABLES ...........................
Chapter
I. INTRODUCTION .......................
Purpose
Importance of the Problem
A Brief Outline
II. INFORMATION ANTICIPATIONS AND
WEALTH REDISTRIBUTION .............
The Net-Creditor-Net Debtor
Studies
The Consumer Sector
Attitudes and Anticipations
Summary and Suggested Integration
III. THE CONSUMER PORTFOLIO ADJUSTMENT
PROCESS IN IMPERFECT MARKETS . . .
The Standard Theory of
Portfolio Selection
Toward an Augmented Standard Theory
Summary
IV. USING SURVEY DATA TO TEST FOR
CONSUMER AWARENESS, ATTITUDES
AND ANTICIPATIONS .................
The Data and Its Specifications
Testing the Hypotheses
Question Set A
Question Set B
Question Set C
Question Set D
V. RESULTS OF TESTS
Tests of the Hypotheses
Chapter
The Chi-Square Tests
Analysis of the Chi-Square Tables
Results of the Probability
Correlation Tests
VI. CONCLUSIONS AND RECOMMENDATIONS........ 133
BIBLIOGRAPHY.................................. 142
iv
LIST OF TABLES
Table Page
1. Portfolio Values of Representative
Men Years 1968, 1969 107
2. Correlations Between Wealth and
Answers (Proportional) to Ques
tions with Significant Chi-Square
Statistics................................125
v
CHAPTER I
INTRODUCTION
Purpose
The purpose of this research is to explore some
of the distributional effects that the execution of mone
tary and fiscal policies have upon different wealth
groups in America. As these policies operate through
affecting the interest rate and level of prices, an at
tempt will be made to show that much of the wealth redis
tribution comes about because of differences in the
levels of information and the ability to anticipate vari
ous market activities affected by changes in these vari
ables. It is expected that a better understanding of
these distributional effects will aid in the assessment
of how government policy operates to affect the well
being of the country.
Importance of the Problem
Understanding the redistribution of wealth gen
erated by government policy is important for two reasons:
first, a proper distribution of wealth is a goal pursued
by the government and the achievement of this goal is
2
dependent in part on the extent to which it is enhanced
or inhibited by the indirect effects of monetary and
fiscal policy; second, a more complete understanding of
the processes through which government policy operates
may be gained from exploring its distributional effects
as they work through the economy. Specifically, it
would be useful to know how differences in information
and anticipations concerning interest rates and the price
level can lead to a redistribution of wealth when govern
ment policy affects the levels of these variables.
The task of attaining the proper distribution of
wealth is commonly attributed to various explicit direct
government policies in the form of tax and subsidy activ
ities such as progressive taxes, welfare, social security,
and so forth. These activities are expected to perform
the major portion of the wealth redistribution. Over the
last twenty-five years these direct policies conducted by
the distributive branch of government (on the fiscal
side) have been leaning toward a more equitable distri
bution of wealth. Questions relating to whether or not
the indirect effects of government policies have become
more equitable have not been adequately considered. The
emphasis of the present study is on the exploration of
these indirect effects.
Interest rates and the price level are affected
3
by government policy tools — buying and selling bonds,
taxes and transfers, government purchases and sales, re
serve requirements and other tools of monetary control.
By altering its policies the government may bring about
a change in either the interest rate, or price level, or
both. This may lead to a redistribution of wealth be
cause individuals in different wealth groups may possess
different amounts of information and/or abilities to
anticipate changes in these variables. This is so be
cause a wealth redistribution may take place when there
are differences between the anticipation of an event and
the actual outcome and differences among individuals or
group anticipations. Since these anticipations are
highly dependent upon the amount of information indi
viduals or groups possess, they both are strongly linked.
Government policy changes are not the only forces
exerted on interest rates and the price level. They are
also affected by numerous real and financial events in a
variety of markets in all sectors of the economy.
Nevertheless, government policies are very important be
cause they require these variables to be at certain
levels as either a direct or indirect consequence of the
policy actions — albeit the semi-artificial levels may
only be maintained for short periods of time.
4
The extent to which government actions must be
carried are determined by the market forces affecting
the variables. In other words, the government reacts to
the numerous real market forces as it carries out its
policies. The more the real forces move toward the de
sired policy levels, the less government action that
needs to be taken and vice versa.
Government policies set up information flows con
cerning the variables they affect. It is often difficult
to detect the degree of government action taken but the
direction is usually evident. Individuals can focus on
these information flows to predict future levels of the
variables they affect. The difficulty and high costs of
trying to assess the other many and varied influences on
the variables is another reason that leads individuals
to use government policy actions as an indicator of how
the interest rate and price level may be expected to be
have in the near future. Hence, the government's poli
cies serve as the cheapest and best indicators of changes
in the interest rate and price level as they respond to
internal changes in the system and also create other
disturbances in the system that affect these variables.1
I
The conclusions drawn from this research are
valid even for situations where changes in the variables
are not generated by government policy actions as long as
they affect the interest rate or the price level, and
5
This research will focus on these differences in
information and anticipations among wealth groups as a
source of wealth redistribution. Higher wealth and
better information are hypothesized to be associated
positively and significantly, and more accurate antici
pations are also hypothesized to be directly related to
higher wealth. It then follows from the first two prop
ositions that those in different wealth groups could be
expected to behave differently with respect to changes
in the interest rate and the price level.
In order to see how wealth is distributed the
utility maximizing process an individual goes through is
explored theoretically. First, the standard theory of
portfolio choice is explored then four real world market
imperfections are added to the analysis. These imper
fections result from differences among individuals in
their: (1) access to information concerning market
events? (2) ability to anticipate future market outcomes;
(3) relative transactions cost; and (4) ability to spread
risk. While the effects of each one are taken up with
respect to individual portfolio decision making and its
relationship to initial wealth holdings, the first two
they set up information flows which are not available at
zero prices.
6
are emphasized. The characteristics of information, how
they are priced and its relationship to time are dis
cussed. A method for determining the optimal amount of
information to be purchased is developed based upon the
individual's initial wealth. This method allows com
parisons of optimal purchase times and amounts for in
dividuals in different wealth groups. These differences
in optimal purchase times among consumers suggest a
dynamic price adjustment process that is a function of
information dissemination.
The hypotheses that wealth and information levels
and wealth and the ability to anticipate correctly are
tested using consumer survey data collected by the Uni
versity of Michigan. Answers to questions pertaining to
recent changes, anticipated changes and behavior toward
changes in interest rates or the price level are em
ployed. Chi-square tests are used to determine which
questions found significant association between wealth
and specific answers. Further tests were performed to
determine the directions and degree of the associations
between wealth and particular answers. These further
tests included an analysis of the rows as well as the
columns of the contingency tables.
7
A Brief Outline
In Chapter II, a brief sketch of the current lit
erature concerning wealth distribution in both the busi
ness and consumer sectors is made. Special attention is
devoted to the importance of anticipations and the
methods used to gauge business and consumer anticipations
of market phenomena. A synthesis of three major ap
proaches taken by researchers in the area will be used
as a foundation for this thesis.
At the outset of Chapter III, the theoretical
aspects of wealth maximization will be discussed and
then augmented to include various market imperfections.
The basis for portfolios to be affected differently as
a consequence of government policy will be explored
focusing primarily upon differences in information and
anticipations among wealth groups. An optimal time and
price for purchases of accesses to information will be
derived along with the relative gains from information
purchases at different times for individuals in different
wealth groups. Finally, a dynamic price adjustment
process that may give rise to wealth redistributions
is sketched.
Chapter IV describes the data used and how it is
organized for this study. It also outlines the major
8
hypotheses of the study and the methods of testing them
using survey questions. Each question's relationship
to the hypotheses is explored and correct answers (where
applicable) are discussed and compared with alternative
responses.
The findings of the tests performed with the
survey data are reported in Chapter V. Test results for
each question are discussed along with the interpreta
tion of these results.
Finally, in Chapter VI, a summary of the findings
and their interpretation, the conclusions and recommen
dations for further research will be covered along with
policy considerations.
CHAPTER II
INFORMATION ANTICIPATIONS AND
WEALTH REDISTRIBUTION
The analysis of wealth distribution has gained
recognition recently as a means of verification and pos
sible augmentation of the usual income distribution
analysis.1 The need for wealth analyses is derived from
the imperfections of various markets which make it dif
ficult for distributional changes to be reflected in an
individual1s income (at least in the short run) . Many of
these changes that do not show up immediately in the in
come analyses are changes in financial variables. Theo
retically, there should be no difference between the
analyses of a stock (wealth) or the flow emanating from
that stock (income).2 However, because of imperfect
markets, this is not so. Income seems to be more sensi
tive to changes in real variables and wealth more respon
sive to financial variables. One major imperfection is
■*Tor a summary of the most predominant neoclas
sical theories of income distribution, see M. Bronfen-
brenner, Income Distribution Theory (Illinois: Aldine
Publishing Company, 1971).
2For an analysis of the stock-flow relationship
in consumer studies, see James Tobin, "The Theory of
10
human capital which is not readily affected by changes in
the interest rate, for example, because it cannot be
traded (i.e., laws do not allow lifetime labor contracts
to be made). This is not to say that there are no imper
fections on the wealth side, but rather that a more
thorough analysis of the distribution of income can be
attained by focusing on the source of the income and the
variables that affect it. For purposes of this study
where the effects of financial variables are of paramount
importance, the distribution of wealth is the most appro
priate course to take.
The analysis of wealth redistribution concen
trates chiefly upon two financial variables — inflation
or changes in the price level and the interest rate.
That is to say, changes (unanticipated) in either or both
of these variables are expected to lead to a change in
the distribution of wealth. While almost all studies of
wealth distribution concentrate directly upon these vari
ables, the authors, in most cases, have at least im
plicitly been thinking about government policy as a major
Portfolio Selection," The Theory of Interest Rates,
ed. by Mann and Brechling (London: Macmillan & Company,
Ltd., 1965).
11
cause of the changes In these financial variables,3
The redistributive effects of unanticipated in
flation are a function of (1) the debtor-creditor status
of the individual (or group), and (2) the degree of price
flexibility of the particular portfolio item held. Net
debtors gain by the amount of unanticipated inflation
times the value of their net debtor position. The de
gree of price flexibility of an item determines how much
an individual will gain or lose from holding that item
during a period of inflation. It is a measure of the
price sensitivity of a portfolio item which includes
those who profit nominally from holding non-fixed mone
tary assets (e.g., stocks, real estate), as well as those
who gain because their human capital is more price
flexible than others.
It should be noted at the outset that all in
vestigators would agree that the redistributional effects
of inflation are all due to the fact that inflation is
unanticipated. If it were anticipated, then the market
prices of all goods affected by the inflation would
^For a study that explicitly recognizes the in
fluence of government policy upon the distribution of
wealth and income, see 0. Brownlee, and A. Conrad,
“Effects Upon the Distribution of Income of a Tight
Money Policy," Commission on Money and Credit, Stabiliza
tion Policies (Englewood Cliffs, N.J.s Prentice-Hall,
Inc., 1963).
12
adjust (sell at premium or discount depending upon the
type of item and the rate of expected inflation) properly
and the question of inflation's redistributional effects
would lose significance.
The Net Creditor-Net Debtor Studies
One of the most common ways of evaluating the
distribution of wealth is to explore the "debtor-creditor
hypothesis" which held that those with a net debtor posi
tion during a period of rising prices will gain at the
expense of the corresponding creditors (to the extent
that the creditors did not anticipate the inflation)
since what they pay back is lower in real terms, due to
the decline in the real value of the means with which
debt is retired. Actually, this is more than just a
hypothesis since inflation is defined as a decline in the
value of money vis-a-vis all other goods. Hence, the
investigators of this hypothesis are not really testing
the "debtor-creditor hypothesis" per se (i.e., that
wealth is distributed from creditor to debtor during
periods of inflation since this is true by definition),
but are trying to determine who anticipates inflation
best.
In an effort to determine whose predictions rate
13
most accurate, a number of researchers have employed the
method of using stock prices as an indicator. This
method incorporates two effects: on the one hand, the
effect that inflation has on the net wealth of the busi
ness which is determined by the net debtor or creditor
status of the firm, and on the other hand, the return
(interest rate) that the firm receives (or pays) for
holding that particular debtor or creditor position. For
example, if a firm is a net creditor during a period
where prices are increasing by (say) 5 percent, then the
value of the firm's net worth will fall by 5 percent of
the value of the creditor position (ceteris paribus).
This, however, is not the whole story as they may be com
pensated by (say) 10 percent for holding this creditor
position which would produce a profit rather than a loss.
Ideally, the stock market price should reflect
both effects. Whether it does or not is difficult to
determine since there are so many variables that influ
ence a stock's price and the sum total of these effects
may reflect different degrees of awareness of these
effects at different times.
One of the first to make use of these kinds of
tests was K e s s e l ^ who used three sets of sample firms
^R. Kessel, "Inflation-Caused Wealth Redistribu
tion: A Test of a Hypothesis," American Economic Review
(March 1956).
14
listed on stock exchanges. He correlated debtor
(creditor) status and inflation (deflation) and found
that increases in stock prices of banks (first set) and
industrial firms (second set) were significantly related
to their net debtor (creditor) status. The test was a
rank correlation and approximately 23 percent of the
variation in stock prices could be explained by the net
debtor status. He also found significant results in the
Mann and Whitney test for significant differences in
random variables for a deflation (1929-1933) for a random
sample of New York Stock Exchange firms and net creditor
status.
Improving and expanding upon Kessel1s study,
Kessel and Alchian5 explored the effects of inflation
upon four different industries. They made use of the
"t“ test to determine if there were significant differ
ences between the mean value of stock prices of net
debtors versus net creditors. Their results showed those
relationships to be significantly different. They dem
onstrated, in fact, that net creditors gained during
periods of deflation and net debtors gained during peri
ods of inflation. One major drawback of this kind of
5R. Kessel, and A. A. Alchian, “Effects of In
flation, " Journal of Political Economy (December 1962).
15
analysis is the assumption that "a firm that was a net
monetary debtor one year was very likely to be one the
next year."6 This, in fact, was not so, and the value of
the debtor or creditor status during the aberrant periods
was not calculated.
Meanwhile, Bach and Ando? examined the debtor-
creditor hypothesis during the inflationary period of
1939-1952. Because of the changing of the net creditor
status (which Kessel and Alchian would not allow at all),
the period was divided into three intervals. The 100
randomly chosen firms' stock prices and returns on in
vestment were ranked and correlated with their net debtor
status (rank correlation being the test used). The re
sults were inconclusive and do not confirm the hypothesis
that net debtors gain in periods of inflation. As far
as sectoral distributions go, they found that the con
sumer sector was a net creditor and that the government
sector a net debtor. Hence, during inflation, the gov
ernment would gain at the consumers' expense.
Formulating the hypothesis in terms of expecta-
6Ibid., p. 537.
7G. L. Bach, and A. Ando, "The Redistribution
Effects of Inflation," Review of Economics and Statistics
(February 1957).
16
tions, DeAlessi® sets up a model that measures the degree
of accuracy of anticipations of inflation. He tests the
model with firms listed on the United Kingdom stock ex
changes. His test of the debtor-creditor hypothesis is
developed in terms of an accuracy of inflation anticipa
tions coefficient. If this coefficient of anticipations
is significantly different from zero, then there has
been a redistribution of income (wealth). (The coeffi
cient represents the difference between actual and
anticipated inflation divided by the actual inflation.)
Fitting a stochastic model, he finds estimates
for the coefficient. These estimates are, however, in
conclusive; the null hypothesis that the coefficient is
equal to zero is not rejected in almost all cases. Al
though his other tests do not offer much more credence
to his hypotheses, DeAlessi holds on the basis of this
evidence that the debtor-creditor hypothesis is not found
to be inconsistent with the data. Hence, business firms
were found not to anticipate inflation correctly during
1948-1957 and accordingly, there was a redistribution
of wealth from creditors to debtors.
L. DeAlessi, "The Redistribution of Wealth by
Inflation: An Empirical Test with United Kingdom Data,"
Southern Economic Journal (October 1963).
17
Prichard^ tests an augmented version of De
Alessi1s model which incorporates both buyers and sellers
of bonds and allows for both an underestimation of infla
tion as well as an overestimation. Eight hundred and
forty-nine firms are sampled during three different in
flationary periods and are tested using a non-parametric
rank correlation, a linear regression and the Wilcoxen
test. He found that there were few discernable cases
where the redistribution of wealth resulted from the co
efficient of anticipations being significantly different
from zero. Hence, Prichard concludes that the infla
tionary induced wealth redistribution is not likely to
occur through errors in the anticipations of inflation.
Adding up the score sheet of findings of the
debtor-creditor hypothesis, the results are inconclusive.
Kessel and Alchian say yes; DeAlessi and Bach and Ando
are not definite with DeAlessi leaning toward a yes; and
Prichard's a staunch no.
While the above studies are admirable attempts
at determining how wealth is distributed in the business
sector, they fall short of their mark. This can be seen
^W. C. Prichard, "Inflation, Expectations and
Wealth Distribution" (unpublished Ph.D. dissertation,
Michigan State University, 1971.
18
by noting that the difference between all the above in
vestigations are slight and involve the choice of data,
duration of the test and, of course, how the data are
aggregated. Perhaps one reason for this is that stock
market prices (which are relied upon heavily in these
analyses) reflect many variables beside net creditor
status. Another possible error is that they neglected
to break the firms down by size. The larger ones might
act very differently than the smaller ones due to vary
ing amounts of market information, ability to spread
risk, etc. These potential sources of errors will be
reduced in the following analysis of the consumer sector
where consumers will be broken down by size of portfolio
holdings.
Finally, even if the results were confirmed by
all the investigators, they would still not necessarily
hold for the consumer sector due to the large differ
ences between consumer and business behavior (e.g., to
ward risk). Hence, an analysis of wealth distribution
would still be required for consumers and the business
sector analyses would not offer much in the way of
assessing consumer behavior (especially since there are
no market prices for individuals). They make no sugges
tions how to measure consumer anticipations, how they
19
are formed, how they are distributed or what effect they
may have on the distribution of wealth in the consumer
sector.
The Consumer Sector
Answers to the above questions require a review
of the consumer sector studies of wealth distribution.
The following are the most relevant:
PeseTc^O makes an attempt at finding out how the
consumer sector is affected by inflation. He breaks the
consumer sector down into income groups and then classi
fies them according to their net creditor status much
like Bach and Ando3-1 did in their analysis of inter
sectoral wealth distributions. Prom this initial net
creditor status (which is assumed fixed), he calculates
the redistribution of wealth that would occur from a
small amount of inflation and alternatively three tax
policies (including a sales tax on food) that would all
raise the same amount of revenue for the government. He
finds that the inflation is the most regressive tax and
B. P. Pesek, "A Comparison of the Distribu
tional Effects of Inflation and Taxation," American
Economic Review (March 1960).
11-Bach and Ando, op. cit.
20
redistributes wealth s i g n i f i c a n t l y . 1 2
Pesek's analysis is shackled by his unreasonable
assumptions that the net creditor status will remain the
same and that markets (influenced by consumers) will make
no adjustment to compensate for inflation (e.g., interest
rates increasing). What he is assuming in effect, is
that consumers adjust very slowly to inflation or are
oblivious to it. This is the very thing DeAlessi^3 and
P r i c h a r d ^ are testing for in their models of the busi
ness sector's response to inflation.
•^Robinson Hollister, and John Palmer, "The Im
pact of Inflation on the Poor," Discussion Paper No. 40-
69 (Institute of Research and Poverty, University of
Wisconsin, 1969); and Albert E. Burger, "Effects of In
flation (1960-1968)," Federal Reserve Bank of St. Louis
(November 1969), pp. 25-36. These authors have shown
that inflation has benefited the poor. But their bene
fits arise not from inflation per se, but from assist
ance programs that have developed after World War II.
Whether these programs are a direct result of inflation
is difficult to determine. However, even if they are,
then Pesek's analysis is still correct as the government
would be trying to offset the extra burden imposed by
inflation on the lower income groups.
1 ^
DeAlessi, op. cit.
^^Prichard, op. cit.
21
Other attempts at determining the incidence of
inflation on the consumer sector include Goldsmith and
Lipsey's15 leverage ratio which measures the potential
capital gain (loss) realized due to price level changes.
It measures the price sensitivity of various portfolio
items and can be used to simulate the effects of price
level changes on different consumer portfolios. Gold
smith and Lipsey find that the size of leverage ratios
are related to a number of socioeconomic variables.
Housing status, income levels, age, and occupations were
the specific variables used in their analysis employing
survey data (data were from Consumer Union and Survey of
Consumer Finances). From this data they simulate an in
flation assuming that the initial holdings remain the
same. The analysis led them to conclude that inflation
hurts renters, blue-collar workers, and the young most.
An offshoot of the Goldsmith and Lipsey analysis
is Projector's^ study in which she computed elasticities
for asset preference with respect to net worth. Building
on Projector's work and Goldsmith and Lipsey's leverage
15
Raymond Goldsmith, and Robert E. Lipsey,
Studies in the National Balance Sheet of the United
States, N.B.E.R. Study (Princeton: Princeton University
Press, 1963).
■^Dorothy Projector, "Consumer Asset Preference,"
American Economic Review (May 1965), pp. 227-251.
ratios, Budd and Seiders-^ have set up a model of the
consumer sector Which indicates how inflation would
effect the distribution of income and wealth. They find
"adjustment coefficients" (estimated from post World War
II data of inflationary effects on the real value of
various portfolio items) which indicate the change in
value of the portfolio items with respect to inflation
and use them to estimate the effects of inflation on dif
ferent income groups holding different items. Data con
cerning income, assets and/or debt items are obtained
from the Survey of Financial Characteristics of Consumers.
Their findings suggest that there is a small but con
sistent movement to greater equality as a result of
inflation when considering changes in net wealth and an
almost negligible change in the distribution of income.
While this article is probably the best analysis
of wealth distribution in the consumer sector so far, it
does not try to explain why different income or wealth
groups hold different kinds of assets (with different
adjustment coefficients) and hence cannot explain struc
tural shifts in the balance sheet of individuals over
time.
■^Edward c. Budd, and David Seiders, "The Impact
of Inflation on the Distribution of Income and Wealth,"
American Economic Review (May 1971), pp. 128-138.
23
Another important study in the analysis of income
and wealth distribution in the consumer sector is
Brownlee and Conrad.Their analysis is much broader
than those discussed above as it tries to determine what
the effect would be on an average income class (with
respect to real income) if there were an increase in the
demand for goods, and (1) if there were no offsetting
monetary of fiscal policy; (2) if government policy were
used to increase the money supply to maintain the price
level; or (3) if there were an increase in taxes or a
decrease in government spending. That is to say, what
would be the effect of tight money, inflation or higher
taxes on the distribution of income? (They assume that
government policy works mainly through changes in the
interest rate which then lead to changes in the distri
bution of income.)
Brownlee and Conrad find that inflation is a
burden to those in income groups below $6,000, and helps
those above that level. Tight money accomplishes exactly
the opposite effect and increases in taxes (of three
types specified) will yield progressive burdens on all
income groups. Their findings are based upon a sample
of 15,000 subscribers to Consumer Reports (a mailed
questionnaire) in 1957. The "coefficient of adjustment"
•^Brownlee and Conrad, op. cit.
24
derived from the sample were then used to measure the
distribution of income in the three simulated experiments
replete with different assumptions (concerning the dis
tribution of corporate income tax to consumer expen
ditures) .
Their analysis is laudable for its vastness of
scope and attempts to measure different effects of dif
ferent policy actions under given constraints. Much of
their results, however, are preliminary rather than con
clusive. Pesek's review^ revealed a potentially sig
nificant omission in their work as they did not calculate
the costs of deflation associated with the tight money
experiments (i.e., only the cost of the interest rate
increasing was calculated). Other errors may reside with
the inadequacy of the data employed and the assumptions
of how government policy works.
One important aspect of the distribution of
wealth that has been somewhat slighted or ignored in the
consumer studies concerns anticipations. DeAlessi,
Prichard, and others, have pointed out that unexpected
inflation or incorrect anticipations will lead to a re
distribution of wealth of firms. Similarly, correct
■^B. P. Pesek, "Macroeconomic Theories of Income
Distribution: Discussion," American Economic Review
(May 1961), pp. 88-91.
25
anticipations may also play an important role in the dis
tribution of wealth in the consumer sector. Changes in
anticipations which have not been included in the analysis
of Pesek,2^ Brownlee and Conrad,ana others, must be
reconsidered if the total distributional effect is to be
taken into account.
For an analysis of how consumer anticipations
effect the distribution of wealth through the net debtor-
creditor position or by any other means where the distri
bution of anticipations is different and potentially
incorrect, some of the consumer anticipations studies
will be reviewed below. In addition, some of the ques
tions asked of consumers relating to wealth distribution
will be used in this analysis to measure consumer
expectations.
Attitudes and Anticipations
The process whereby adjustments for inflation or
changes in interest rates are made by consumers is not an
automatic one, and it does not seem to follow any one
specific pattern. In order to detect some of these pat
terns of adjustment, various researchers have relied upon
20pesek, "A Comparison of the Distributional
Effects of Inflation and Taxation."
21Brownlee and Conrad, op. cit.
26
consumer attitudes toward inflation, interest changes,
and related phenomena. They find that the measurement
of consumer attitudes can be a powerful predictive tool
of future consumer behavior because, until the individual
receives enough information about a particular event
(which will then lead him to change his anticipations),
he will not change his behavior despite the fact that
the actual market event has occurred. The process be
comes more complex when learning new relationships are
required of the individual, and hence there may be long
lags before adjustment takes place to a change in a mar
ket parameter. For example, Katona^2 has shown that
almost all individuals are able to gauge the effects of
changes in income taxes upon their portfolio and will
adjust their portfolio with only a small delay. However,
inflation's effects have not been adjusted (even though
they may have the same effect upon their wealth) because
consumers were not familiar with their potential effects
and, furthermore, could not be sure that the inflation
would persist, whereas taxes persist until the next tax
law is introduced.
It is not surprising that researchers in this
^^George Katona, "Attitudes Toward Fiscal and
Monetary Policy," Public Policy (Winter 1970).
27
field hold that these anticipations are an important
element in the determination of consumer behavior be
cause there are "many a slip twixt cup and lip," i.e.,
between an event and its outcomes. Hence, the process
of converting consumer attitudes into behavior must be
considered. According to M u e l l e r , it is as follows:
first, some sort of stimulus affects the usual markets
(say, inflation); second, consumers perceive various
trends in the market (prices increasing); third, from
these perceptions, an attitude is formed which finally
results in a particular action which is different from
what would have occurred if the attitude had not been
formed. Thus, the intervening variables separating the
stimulus and response must be identified and measured in
order to gain a better predictive tool for consumer be
havior. 24
23Eva Mueller, "Consumer Reactions to Inflation,"
Quarterly Journal of Economics (May 1959), pp. 246-262.
^The importance of observing and understanding
consumer behavior has been proffered by: George Katona,
"Changes in Consumer Expectations and Their Origin," The
Quality and Significance of Anticipations Data, N.B.E.R.
(Princeton: Princeton University Press, 1960), pp. 53-90;
"Expectations and Decisions in Economic Behavior," The
Policy Sciences, ed. by D. Lerner and H. Lasswell (Stan
ford: Stanford University Press, 1951); Psychological
Analysis of Economic Behavior (New York: McGraw-Hill
Book Company, 1951) ; and Eva Mueller, "Consumer Atti
tudes: Their Influence and Forecasting Value," The
Quality and Significance of Anticipations Data. N.B.E.R.
28
The main thrust of these analyses of attitudes
and consumer behavior are directed toward consumer pur
chases (usually of durables) . Changes in portfolios are
indirectly dealt with and their effects on wealth dis
tribution can be found mostly between the lines.
Examples of the latter include Katona and
Mueller2^ who consider forms of discretionary savings in
a two-year re-interview survey of consumers conducted in
1950-1952. The respondents were broken down into three
income groups and their portfolios into liquid assets and
private investments. Questions relating to why people
held different forms of assets were asked but unfortu
nately little else was done with the data.
Mueller2^ found a significant relationship be
tween long-run increases in prices and increases in the
demand for real estate and stocks in 1958 (the demand for
these assets is derived from answers to questions con
cerning attitudes toward inflation).
(Princeton: Princeton University Press, 1964); and "Ef
fects of Consumer Attitudes on Purchases," American
Economic Review (December 1957), pp. 946-965.
25George Katona, and Eva Mueller, Consumer Atti
tudes and Demand 1950-1952 (Ann Arbor: University of
Michigan, 1951) .
26Mueller, "Consumer Reactions to Inflation,"
pp. 246-262.
29
Katona,27 on the other hand, finds that only a
small percentage of large asset holders switch from one
savings medium to another because of interest rate dif
ferentials. Prom this, he concludes that consumers deem
savings media to be insensitive to interest rate changes.
Nevertheless, he finds that interest rates are signifi
cant for housing purchases but not for ordinary durables
or non-durables. (Again, all demand estimates are de
rived from answers to questions.)
Other studies28 have drawn attention to the im
portance of gauging consumer attitudes and expectations
in order to better predict their behavior by demon
strating that an event's occurrence does not lead to a
change in consumers1 behavior until their attitudes
change. They have also shown that different consumers
held different forms of assets and that different con
sumers had varied anticipations. However, the question
of how these relationships fit together was not explored.
^Katona, "Attitudes Toward Fiscal and Monetary
Policy."
^®Mueller, "Consumers' Attitudes Toward Saving and
Their Investment Preferences"; Eva Mueller, "Ten Years of
Consumer Attitude Surveys; Their Forecasting Record,"
Journal of the American Statistical Society (December
1963), pp. 899-918; Katona, "Expectations and Decisions
in Economic Behavior"; and "Attitudes Toward Fiscal and
Monetary Policy"; and John Lansing, and S. Withey, "Con
sumer Anticipations; Their Use in Forecasting Consumer
Behavior," Studies in Income and Wealth, Vol. 14, N.B.E.R.
30
Summary- and Suggested Integration
It has been shown that a redistribution of wealth
cannot occur without there being incorrect anticipations.
This was what DeAlessi and Prichard had in mind when they
made their analyses of the business sector. In the con
sumer sector anticipations investigations, however,
consumer attitudes and expectations have been assumed
away or ignored. The advocates of consumer attitude sur
veys, on the other hand, have focussed directly upon
anticipations and attitudes of consumers but have not in
quired into the wealth distributional effects of diver
gent attitudes.
The ostensible differences (or independence) of
these approaches are clearly demonstrated in a comprehen
sive theoretical model of consumer maximizing behavior
constructed by Miller and Watts. ^9 Their main model
yields the value of a person's portfolio over a lifetime
(or at any one point in his life) . The cash flow from
this portfolio value is "implicitly dependent upon his
[the consumer's] experience and information and explic
itly dependent upon his cumulative lifetime amount of
(Princeton: Princeton University Press, 1955) .
^Robert F. Miller, and Harold W, Watts, "A Model
of Household Investment in Financial Assets," Determi
nants of Investment Behavior, ed. by Robert Ferber,
31
effort at portfolio management. . . ."30 in other words,
they explicitly recognize the importance of expectations
and information to portfolio management but when it comes
to dealing with it, they do not look to the consumer
anticipations or knowledge of the particular markets in
volved, but to a set of roundabout and perhaps inaccurate
measures of them.
In order to reconcile the efforts of those inves
tigators who have taken the different approaches men
tioned above, this thesis will attempt to integrate the
anticipations of the consumer with wealth redistribution
in the consumer sector.
The researchers who investigated the business
sector point out the importance of anticipations. How
ever, they did not disaggregate enough to get significant
results and they did not leave much for the investigators
of wealth distributions in the consumer sector. On the
other hand, in analyzing the consumer sector most re
searchers arbitrarily disaggregated on the basis of
income and played down the importance of anticipations.
N.B.E.R. (New Yorks Columbia University Press, 1967),
pp. 357-380.
30Ibid., p. 364.
32
In this analysis, the consumer sector will be explored
on the basis of disaggregate wealth groups (by deciles)
employing the information collected by the Survey of Con
sumer Finances which reflect more clearly (and more
quickly) the effects of changes in government policy
upon an individual's portfolio. The importance of an
ticipations will be taken into account by using more of
the Survey data which makes inquiries about individuals'
knowledge of markets and their ability to anticipate
future market activities and levels. The importance of
anticipations and information in the distribution of
wealth will be developed theoretically in Chapter III.
CHAPTER III
THE CONSUMER PORTFOLIO ADJUSTMENT
PROCESS IN IMPERFECT MARKETS
As stated at the outset, the main purpose of this
research is to identify the redistribution of wealth that
takes place within the consumer sector as a result of
changes in certain market variables which are generally
regarded to be a function of government policy. The
usual way of approaching the problem is to treat wealth
redistribution in a simple but perhaps perfunctory man
ner. The redistribution of wealth that takes place as a
result of government policy is postulated as being deter
mined simply by the form in which an individual decides
to hold his wealth (i.e., what particular items comprise
his portfolio). For example, those holding bonds when
the interest rate increases or cash when the price level
rises will experience real losses while those who hold
large amounts of debt when prices are rising will realize
gains in the real value of their portfolio (assuming that
the change in market prices was not anticipated and re
flected in the interest rate). In this fashion govern
ment policies, via their effects upon various financial
33
34
and real variables, distribute wealth by affecting the
specific components of the individual's portfolio. This
method while neat is a bit too simple. It yields a pic
ture of what "would" happen to an individual or group
holding a particular portfolio if one or more market
prices were to vary, but is shackled by the ceteris pari
bus set of conditions that do not allow rearrangements
of the portfolio as those market prices begin to change.
Implicitly it requires that individuals either adjust
very slowly to market price changes or are virtually ob
livious to them and hence do not (or cannot) make any
alterations to compensate.
Aside from the obvious drawbacks of this simplis
tic method of measuring wealth distributions, a lacuna
remains: why do people hold assets and liabilities in a
particular form? The answer to this query is the popu
lar response that people attempt to maximize their wealth
(as measured in terms of utility) . However, the answer
^-An alternative theory holds that anticipated
price changes will be met with immediate changes in the
market prices which are caused by an immediate shift in
portfolio composition. See R. Kessel, and A. A. Alchian,
"Effects of Inflation," Journal of Political Economy
(December 1962), pp. 521-537. This opposite extreme
rules out any kind of dynamic portfolio adjustment proc
ess or price adjustment mechanism. The problem with
this kind of analysis will be discussed later on in the
section.
35
does not go far enough in its general state because if
people are wealth maximizers then why is it not everyone
in a similar risk category adjusts his portfolio every
time interest rates change significantly or are antici
pated to change?
Before that question can be answered the standard
theory of the portfolio adjustment process, which is a
function of the wealth maximizing process must be ex
plored in order to see which variables are of signifi
cance and what modification may be necessary to bring it
closer to observed portfolio selection processes.
The Standard Theory of
Portfolio Selection
Modern portfolio selection theory, according to
Markowitz^ and Tobin, ^ focuses on two characteristics
that relate to an individual's entire portfolio as well
as to each element in it, namely, risk and returns, in a
wealth maximizing model. Within this framework, the form
which an individual decides to hold his wealth is a func-
^Harry M. Markowitz, Portfolio Selection; Effi
cient Diversification of Investments (New York: John
Wiley & Sons, Inc., 1959).
^James Tobin, "The Theory of Portfolio Selec
tion, " The Theory of Interest Rates, ed. by Mann and
Brechling (London: Macmillan & Co., Ltd., 1965).
36
tion of his desire to maximize the utility from his port
folio selections. Higher expected returns add to the
individual's utility while higher risks (for the same
expected return) detract from it. This may be shown by:
U = f [E(a), VAR(a)] (l)
where E(a) is the expected return (positive for assets
and negative for liabilities) from portfolio items and
VAR(a) is the risk involved as measured by the variance
of the portfolio item's price changes (weighted).
In the selection of the items to be included in
his portfolio the individual can find a set of "effi
cient" combinations of items that will yield the highest
expected return at the lowest possible accompanying risk.
Prom this efficient menu of risk and returns, he may then
choose the one combination that will maximize his utility
subject to his preference for risk.
For a simple demonstration of how an efficient
locus of portfolio combinations is determined, a two-
asset portfolio is employed.4 Consider two assets, A and
B, with their respective risk-return combinations at A
4This may be generalized into larger portfolios
through the use of a quadratic programming approach
developed by Markowitz, op. cit. See also Eugene M.
Lerner, and Williard T. Carleton, A Theory of Financial
Analysis (New York: Harcourt, Brace & World, 1966).
37
and B on the following graph:
GRAPH I
. . B
Return
(Ea)
If the expected risk and return outcomes from each asset
were perfectly correlated with each other, the locus of
portfolio combinations would be a straight line from A
to B. But if some of the outcomes are independent and/or
negatively correlated, the total portfolio risk will be
smaller than the average individual risk so that the
combination of total portfolio risk-return points might
be ACB. In the extreme case where A and B were perfectly
negatively correlated, the risk can become zero for a
specific combination and the locus of portfolio combina
tions would then become ADB. Only DB is "efficient,"
however, because at each point along AD there is a cor
responding point along DB which offers a higher return
for the same amount of risk.
VAR (a)
38
The efficiency locus can be derived mathemati
cally. For any E(a) a specific combination of A and B
can be found by:
E ^ ) = Wx E(a1) + W2 E(a2) i = 1,2. (2)
where the W s are weights of asset combinations whose sum
adds to one. The variance for this combination of A and
B can be found by the following equation (determined by
the calculus of expectations) : ^
VAR(W^ax, W2a2) = wfvAR(ax+ W2VAR(a2) + 2 WXW2 COV(
(3)
The covariance (COV) between the two assets indicates how
the risk on the assets are related. Obviously, any nega
tive relation (correlation) would reduce the total vari
ance of the portfolio since the COV would be negative
making the third element in (3) negative thus reducing
the total variance. Furthermore, even positive covari
ances will reduce the total variance below that of either
item alone as long as the covariance is less than the
individual variance of each item. Thus, total portfolio
5When three or more items are involved, various
combinations of these items may yield the same E(ai) but
different variances, quadratic programming can be used
to find the optimal choice of assets where the E(aj.) is
highest for a given VAR(ai) . See Markowitz, op. cit.
39
risk can be reduced as long as the covariances are nega
tive or smaller than the variance of either asset.
The next step in the analysis is to find the in
dividuals' preference for risk. These preferences are
reflected in the indifference curves below (Graph II).
Higher indifference curves are found to the right
(12 ^ 1^) as they yield higher returns at each level of
risk. The slope of the indifference curve (I) at any
point represents the individual1s marginal rate of sub-
stitution of risk for return.
^The marginal rate of substitution of risk for re
turn can be found as follows: Given U = f [E(a), VAR(a) ]
for each indifference curve I, XJ is some constant.
Setting the function equal to a constant and taking the
whole derivative we get
du d E(a) + d VAR(a) = C
dE( a) dVAR(a)
dU ... 3U , d E(a)
where a¥(iT>0> aVARl' a)' and solving for '3~VAR(aT?
9 U
9 £ (H i )
we find — --- which is the marginal rate of substi-
9 VAR ( a)
tution of risk for reward. Diminishing marginal utility
(or in this case increasing marginal disutility) re
quires that the second derivative be negative or
_ a 2u + 2 »iu + ._*i2-- < o.
dE(a)2 * 9 E(a)2 dE(a) 0VAR(a) aVAR(a)2
40
GRAPH II
VAR (a)
E(a)
The marginal rate of substitution of risk for return is
depicted above as a decreasing function. This is due to
the individual's desire to obtain proportionally larger
amounts of return in order to be induced to incur more
risk. That is to say, risk averters experience increas
ing marginal utility) when trading returns for risk.
In order to maximize his wealth, the consumer
will choose the point of tangency between his utility
curve and the efficiency locus of portfolio combinations
by moving along the E-E curve (efficiency locus; in
Graph I, CB or DB) to the highest utility curve. In
other words, he will choose that combination of port
folio items that will yield the highest utility given his
preference for risk and his expectations of the particu
lar items market performance. This can be shown by
41
superimposing Graph I on Graph II as shown in the illus
tration of Graph III.
GRAPH III
VAR(a)
E(a)
In this manner the standard theory explains an
individual1s behavior with respect to portfolio selec
tion. As indicated earlier, there still remain unan
swered questions by the analysis. For example, why does
not everyone adjust their portfolios at the same time and
in the same direction (given the same degree of risk
aversion) , or more fundamentally, why does not everyone
with the same risk preference have the same proportional
portfolio? The answers to these kinds of questions will
be taken up in the following section.
42
Toward an Augmented Standard Theory
The approach to augmenting the standard theory
will involve three important influences that will have
to be considered and incorporated into the analysis of
portfolio selection. These influences are: (1) trans
actions costs and capital market imperfections; (2) the
information problem; and (3) the ability to spread risk.
It is hypothesized that these three influences have a
significant bearing upon the decision to adjust a port
folio and may therefore explain why adjustments and
selections are not the same among individuals with
identical risk preferences.
A. Transactions Costs and
Market Imperfections
The efficiency locus derived in Graph I needs to
be augmented if transactions costs are to be included in
the analysis of portfolio selection. This is necessary
because the purchase and/or sale of any asset or lia
bility usually requires the services of a salesman or
broker. Charges for these services may be a fixed amount,
some function of the size of the transaction or both.
An individual seeking to change the composition of his
portfolio will have to calculate the costs involved in
43
making the necessary transactions. Thus, the expected
return from the portfolio will have to be reduced by the
amount of the costs to make the transactions, or
E(ai = Eai - (ax + 0 iplal) + e^2 - («2 +/* 2P2a2^ +
... + Eam - («m + / » mpmam) (4)
where a is a fixed transations charge, fi is the variable
expense and p^ a^ is the total value of the transaction.
The inclusion of transactions costs will move
the E-E schedule to the left and force the individual
onto a lower utility curve than he would have otherwise
been on if there were no transactions .costs.
If the 1s are positive and the 1s constant
then portfolio adjustments would be less costly to the
persons who carry out large transactions since the fixed
costs can be spread over a larger number of units. In
general it would seem that those making the largest
(absolute) portfolio adjustments would be those with the
largest portfolios.7 in times of varying government
policy an individual with a larger portfolio would be in
a better position to change his portfolio because of the
lower average transactions cost.
7A smaller portfolio holder might make large,
trades also, but he would be sacrificing the risk reduc
tion of diversification by doing so. See Section C on
Spreading of Risk in this chapter.
44
If 0 is not constant then the advantage of the
larger portfolio holders is not clear. Two possible
cases shall be distinguished.
The first possible case is a progressive marginal
cost which would hurt the larger portfolio trader rela
tive to the smaller owners. The second, a regressive
marginal cost structure, would benefit the wealthier man
relatively. For the most part, the latter case and the
constant 0 base prevail and thus the larger portfolio
traders are relatively better off with respect to costs
of making transactions. That is to say, the E-E curve
for the larger portfolio holder would not move as far to
the left when transactions costs are added as that of
the person who does a smaller volume of transactions.
Somewhat related to the transactions cost anal
ysis is the problem of capital market imperfections as
manifested in the indivisibility of various assets and
liabilities. For example, when one person buys a house
another (in a lower wealth group) cannot purchase a frac
tion of a house and thus may be forced to hold some
other asset that may involve a different risk-reward
relationship. This is another reason the E-E curve for
different wealth groups may not be the same. The problem
of not being able to afford to enter a particular market
45
because of indivisibilities would probably be more common
among the lower wealth groups.
B. Information
The second influence to be integrated into the
standard analysis is information. Individuals who
possess varying degrees of information (i.e., quantities)
will also be viewing different E-E curves. This occurs
for the following reasons. First of all, the individuals
in different wealth groups will be viewing efficiency
curves comprised of different assets since the person
with greater information will be aware of more ways of
holding wealth that are more profitable and will be elim
inating those means which are not as rewarding. Second,
even if they are looking at the same assets, those with
the most information will be evaluating the return and
risk on these assets differently than those with less
information. In this way information plays a dominant
role in the process of portfolio selection.
Before discussing the process whereby informa
tion may influence the wealth maximizing adjustment
process, a few words are in order concerning the nature
of information and the measurement of its character
istics.
46
1. Characteristics of Information and Their Meas
urement :
a. Temporal Types and Sources of Information:
Information may he of two temporal types — past
or future. That is to say, an individual may receive
information concerning what has already taken place or
what is about to occur. Information concerning past
events is not always accurate because it is not always
clear what has transpired irrespective of the set of ex
planatory variables used to describe the event(s). For
example, it took over a year for most businessmen and
economists to recognize a recent recession. On the other
hand, information concerning the future is rarely per
fectly accurate. This is why the individual cannot
accept a piece of information as being correct; instead,
he makes probability assessments (subjective) concerning
each piece of information he receives.
When he receives the information the individual
must make two probability assessments. The first con
cerns the correctness of the information received. The
second judgment relates to the effects of the information
(expected) upon different markets.
How reliable a piece of information is, depends
primarily upon its source. There are three basic
47
sources of information: (1) market information, (2) in
side information, and (3) rumors.
The first source of information is market infor
mation which is available to anyone willing to pay the
price for it (even if the price is buying and reading a
newspaper) . While general information is made available
to many people relatively inexpensively, specific infor
mation, concerning one particular market for example, is
more expensive, and the more accurate the information
(average) the more reliable the source and the more ex
pensive the information.
Inside information may be defined as information
that is available only to a select group concerning the
operations of markets that some of the "insiders" either
have control over or have more information than is cur
rently available to everyone (i.e., available to be
purchased in the normal market for information) .
Finally, rumors are similar to inside information
except for two features: first, there can be no control
over market forces by the informant, and, second, the
accuracy of prediction is much lower than inside informa
tion (i.e., the probability of realizing the expecta
tion) .
48
b. Pricing Information's Characteristics:
Theoretically, these sources of information can
be differentiated. In actual practice, however, few
people can point to a source for a particular piece of
information and, even if they could, the problem of
classifying reliability within these groups would still
exist (e.g., one piece of inside information may be more
reliable than another). Perhaps the simplest way of
handling this problem is to use the price of the informa
tion as a proxy variable for how reliable the information
is. In this way, all sources of information could be
classified immediately and would in almost all cases fit
in with the three-way division of sources outlined above,
i.e., from least expensive to most expensive, rumors,
market information and inside information.
After the information's reliability has been
assessed another probability judgment must be made,
namely, what kinds of effects will result from the action
or change (real or expected) that the information re
ports. (For example, substitutes, if it is used as a
factor, e.g., in production of hogs, and so on.)
The second assessment relies upon the kinds and
accuracy of associations that can be made between the
information received and the potentially affected mar-
49
kets. It seems plausible that the individual with the
most information (on the average) will be the one to make
the most associations and that his will be the most accu
rate ones. Exceptions to this theory might be those with
greater education or skills in dealing with the specific
markets affected by the information. However, education
is not free and so information processing and manipu
lating may also be considered priced (albeit in different
markets). As for special skills taught only in certain
industries, which may allow for more accurate associa
tions, the value of developing this skill should be dis
counted by the employer of the individual and his wages
reduced by that amount (otherwise, there would be an ex
cess supply of workers for these jobs due to the higher
"effective wage" which would tend to reduce the actual
paid wage). Hence from a theoretical point of view,
information is priced irrespective of source or temporal
variety.
The price of information, then, represents the
reliability of information as well as the accuracy and
number of associations that can be derived from the mar
ket data. While there may be many drawbacks in placing
all bets on one variable that is subject to many market
imperfections, it is extremely convenient to work with
50
for purposes of analysis and is not expected to be any
worse than the other more complicated alternatives.
c. The Timing of Information:
One additional feature of information, perhaps
the most important that price is associated with, is when
the information is made available. That is to say, the
value of the information (in terms of reliability, and
so forth) is a function of both price and time. The
price of the information and the time when it is re
ceived are related in an inverse fashion, i.e., the
smaller the time lag between when the piece of informa
tion is first generated and when the purchaser receives
it the higher the price. Thus individuals are thought
of as paying for not only reliable, accurate, and so
forth, information but also early information. Accord
ingly, all values of informational characteristics are
relative to time.
In discussing a market for information over time
it is more convenient to speak of a market for access to
information rather than for bits of information, due to
the ephemeral (and sometimes sporadic) nature of informa
tion. This is a bit more realistic as people will sub
scribe to a service for a year, or hire a research team
51
or call a broker everyday rather than try to seek out one
piece of information at a time. In fact seeking out in
formation piece by piece would be very difficult for
them since they would not know what information to look
for in the first place.
2. Information and the Individual's Wealth
Maximizing Process:
a. Information and the Efficiency Curve:
As mentioned above, information is an important
influence on the efficiency locus relating risk and re
turn. In fact, the access to information may be consid
ered an asset in the portfolio of an individual. Hence,
the E-E curve will represent different combinations of
assets and liabilities plus the expected benefits de
rived from having additional information on how these
assets and liabilities may fluctuate with respect to
their market prices.
When information purchases are included in the
analysis two effects must be accounted for. The first
is the loss of returns on assets that could have been
appreciated if another asset other than information were
held (i.e., the opportunity cost). The second is the
additional benefits that accrue from being aware of dif
ferent ways to hold wealth and those that are the most
profitable (highest expected return), with the least risk.
The first effect tends to move the E-E curve to the left
and the second to the right. The net effect must be to
move it to the right, or else additional information
would not be demanded. However, this does not mean that
all individuals purchase the same amount of information
access. In fact, a marginal purchase of information may
move one wealth group's efficiency locus further to the
right than another1s.
This can be demonstrated most easily by using
average costs of information defined as total costs of
information divided by the number of components in the
portfolio (measured by the dollar size of the portfolio).
Those with the largest portfolios can purchase informa
tion at a lower per unit cost and will therefore purchase
more information than those with smaller holdings who
cannot spread the cost of the information over as many
items. The less wealthy man will be forced to stop pur
chasing information because the per unit cost is greater
t
than the additional return he can get from his portfolio
(expected) as a result of the information. The larger
portfolio owner can spread the cost of the information
over a larger number of assets or conversely has a larger
base over which he can spread the benefits of the informa
tion.
53
b. Optimizing Information Purchases:
The optimum amount of information access to be
purchased by any group or individual then becomes that
amount where an additional piece of information would not
move the E-E curve to the right. In order to determine
this optimum level, the ephemeral nature of information
must be taken into account. A piece of information's
usefulness may last only a day or week or even a mere few
hours. Furthermore, the value of this information is ex
pected to decrease over time. The earlier the informa
tion is available, the greater its worth to the holder as
he has more time to adjust his portfolio to make use of
the effects of the information. As mentioned above,
pieces of information are not purchased separately. What
is purchased is the access to it through newspapers, re
ports, investigations, and so forth. This access to
information has a positive price in the market. If no
information is purchased the price of it approaches zero
but only after a long period of time where the value of
the potential usefulness also approaches zero. The com
bination of information sources an individual subscribes
to can be thought of as an indication of the average
time lag of market information (i.e., the market informa
tion the individual is interested in) he is willing to
accept. The greater his access to information, the
54
sooner he obtains the information (on the average) .
Thus, the value or revenue derived from information ac
cess is a decreasing function of time as shown in the
following Graph:
GRAPH IV
Time
The vertical axis represents the net actual or expected
revenue that would be realized by selling or buying par
ticular portfolio items. As tQ goes to tm the ability
of the individual to adjust his portfolio becomes more
difficult and more costly, because as more and more
people also begin adjusting their portfolios (as the in
formation spreads through the community) , the marginal-
gain realized from the potential transaction (expected
price less the actual price) falls and thus the revenue
from the transaction also falls. The form in which the
revenue curve falls is not important as long as it is
continuous.
55
In order to find an optimum purchase price and
time a cost curve must be introduced. The cost curve is
also plotted against time, as follows:
GRAPH V
Time
The cost curve's shape is a downward sloping
function of time. This is due to the additional costs
that must be incurred to get information early. The
earlier the information is acquired the higher the price
of access. The slight convex shape represents very high
prices of information at early stages and information
having a positive price at all times.
Graphing both the cost and revenue functions to
gether, an optimum can be found where the difference be
tween the revenue and costs are the greatest or at point
te below.
56
GRAPH VI
Cost and
Revenue
t0 te tm Time
When individuals with different size portfolios
are introduced, the optimal purchase time and amount of
profit from information purchases will usually he differ
ent. The most obvious case is one in which the total
costs of information are greater than revenue at each
point in time for one individual (Ri) as in Graph VII.
GRAPH VII
Costs
Revenue
tl t2
Time
57
Individual Rj would not purchase any information in this
case as it would result in negative returns to him, ex
cept at time t2, where he would just break even (zero
profit). Individual R2, on the other hand, would incur
profits of AB or CD if he purchased OB information at
time ti, the optimal purchase time for him.
The revenue curves have different heights (from
the origin) because of the differences in the portfolio
size. Hence, the difference between revenue curves
height is a function of the size difference of the port
folios represented.
Another difference between Ri and R2 is their
slopes. Its importance as far as optimal purchasing time
can be shown in the following Graph:
GRAPH VIII
Total Costs
Total Revenue
Time
58
The optimal purchase time for R2 is t^ where for
Rl it is at a later time, t2- This is due to the steeper
slope of R2 which yields an earlier equality of slopes
(revenue and cost). The profits for R2 are also greater
than for Ri (i.e., EG > PH), at their respective optimal
purchase times. The slope is steeper for R2 at all
points because each unit of time is worth more to the
person with the larger portfolio since he can spread
benefit of that information over a larger amount of
items. This can be shown as follows:
GRAPH IX
Figure (a)
20
Figure (b) Time
In figure (a), the change in t is found to be worth the
same to both Ri and R2 , i.e., $5.00. In figure (b), the
change is $5.00 for R^ and $20.00 for R2. In order to
get a situation where the returns are equal as in figure
(a) for each time unit (i.e., marginal time period) , the
same amount of variable portfolio items must be available
59
to both and R2. In other words, the difference in
portfolio size between Ri and R2 must be in a frozen
form which R2 cannot change in light of new information
(and will not be affected by the new information) . Under
these conditions information would be worth the same to
Rl and R2 at each point in time.
Figure (b) then is the more feasible one since
the assumption of fixed portfolio items unaffected by
information need not be relied upon. In cases such as
these, each time unit would be more valuable to R2 than
Rl. This means that R2 is not only greater than Ri at
all points so that the profit is greater for R2 but also
A Ro
that — - > =■ which is an incentive for Ro to purchase
A t A t
the information at an earlier time since time is more
valuable to him.
In the above analysis, a convex cost function was
employed for the aforementioned reasons. In order to see
how profit maximization may differ for different size
portfolio holders, other continuous shapes will be con
sidered. However, instead of using the analysis where
benefits are spread over a portfolio given a particular
cost function, it turns out to be simpler if the other
side of the coin is explored, namely, the spreading of
60
costs over a portfolio given a specific revenue function.
For this analysis the cost function must be divided on a
per item basis, i.e., total costs at each unit of time
will be divided by the dollar size of the portfolio.
This division is required so that total costs can be
spread over the portfolio just as the benefits were
spread in the preceding analysis. Hence, the optimiza
tion of profits from "per unit" cost and revenue func
tions is essentially the same as maximizing profits from
total revenue and costs. The averaging done in this anal
ysis is for the purpose of finding a common denominator
for different wealth groups' cost and revenue functions
so that they can be compared. Hence, it is still the
total revenue less the total cost per item which is
maximized.
When the revenue curves of Ri and R2 are divided
by the size of the portfolio, they become the same per
item curve as shown in Graph X.
GRAPH X
Cost Revenue
Size of
Portfolio
(dollars)
Time
61
Now, however, the cost functions are different.
C2 corresponds to R2 and to R^. This indicates that
the individuals with the largest portfolios have the
lower cost (per time). The optimizing techniques are the
same in this analysis (as the earlier one) and profits
are maximized for C2 at t^ and for Cj at t2. The earlier
maximization for the larger portfolio owner (C2), and the
larger profit for him also corresponds to the analysis
in Graph VIII.
As mentioned, the convex shape of the cost func
tion is used to represent high initial costs and a posi
tive price of information at all times. Other possi
bilities would include parallel revenue and cost func
tions. While possible, it would be unusual as cost
curves and revenue curves are frequently assumed to be
independent of each other.
A third possibility would be concave cost curves.
With respect to the timing of information purchases, two
possible situations may arise. They are shown in Graph
XI. In (a) profits are maximized at t^ for C2 and ti for
Ci. This assumes that AB < Dti and AC < Dti. If these
inequalities do not hold then both maximize profits at
to. As in case (b), however, the profits are still
larger for C2 even though both Ci and C2 purchase the
62
GRAPH XI
Cost Revenue
-i- Size of
Portfolio
(dollars)
Time
Figure (a)
63
information at the same time. Simultaneous purchases
would also occur (at tQ) if both had linear cost curves
that converged at the T axis (or were close to con
verging) . That is:
GRAPH XII
Time
Cost Revenue
-f- Size of
Portfolio
(Dollars)
tQ would be the optimum purchase time as long as AB and
AC >Dti (profits would again be greater for the man with
the larger portfolio).
Once the slopes of costs curves to different indi
viduals are allowed to vary significantly, fes in the case
above), the following feasible situations may arise.
One case requires that the less wealthy man (C^) purchase
information ahead of the wealthier (C2) man if for some
reason the times when the costs of information become
zero are different (i.e., tj_ ^ t2> as in Graph XIII.
In this case, C2 (with the lower cost curve) purchases
information at ti while C^ buys it at t0 (but even in
this extreme case profits for the wealthier man are
greater than for the poorer man). Significant differ-
64
ences in the slopes and shapes of different groups1 cost
curves would imply that different groups received differ
ent types of information and relied on different sources
for the information, which would be expected.
GRAPH XIII
Cost Revenue
-f- Size of
Portfolio
(Dollars)
t0 tl 12 Time
Even though extreme cases such as the one above
where the larger portfolio holder purchases information
later than the smaller portfolio owner are possible,
these cases are expected to be few and far between. The
reason for this is that the absolute cost of reliable,
and so forth, information screens out those who cannot
spread the costs over large portfolios, and the proba
bility of getting information as early, as reliable, and
so forth, from say, the newspaper as opposed to an in
formation service is not very great since the individual
who has the larger portfolio gets both accesses to the
information. (Even the great newspaper readers like
Will Rogers and I. F. Stone would have a great deal of
trouble getting all the information they wanted, i.e.,
65
accurate, reliable, and so forth, out of reading just one
newspaper, not to mention their influential friends who
also supplied them with news.)
3. The Information Optimizing Process and Price
Adjustments:
Implicit in the derivation of the revenue curve
above is a theory of price adjustment. It is a dynamic
theory where lags exist in the dissemination of informa
tion to the market and prices do not make immediate
adjustments to new equilibrium. The price adjustment
process must be made explicit not only because there are
those who would hold that prices adjust at once, i.e.,
as soon as anyone has the information,® but also because
it is essential in an analysis of the manner in which
a distribution of wealth may take place.
The process through which prices adjust from
equilibrium to equilibrium remains a muddled complexity
to which a great deal of attention has been recently
devoted.^ The idea of information in the market playing
®Kessel and Alchian, op. cit.
^Donald F. Gordon, and Allan Hynes, "On the
Theory of Price Dynamics," Microeconomic Foundations of
Employment and Inflation Theory, ed. by Edmund Phelps,
et al. (New York: W. W. Norton & Company, 1970).
66
an important role in the determination of prices is not
new. It is epitomized in the words of Gordon and Hynes:
"A theory of price dynamics is therefore in reality a
theory of learning.1 1 *0 This learning about markets by
the participants not only determines how prices settle
to new levels but also who is affected by these price
adjustments.
While there is no well developed theory of dy
namic price adjustments available, the static "non-stop"
price theory can be shown to be inappropriate for the
following reasons.
a. Institutional Forces:
First, there are a large number of institutional
constraints in operation that would not allow any one
person to change the value of an asset significantly
(barring the extreme situation where the market is ex
tremely shallow or the extreme assumption of one man
having infinite wealth) even if that individual had per
fect information that the asset price will change by a
significant amount. The reason is that the individual
would have to sell off his other non-liquid assets
and/or borrow in order to take advantage of this informa
-^Ibid. , p. 377.
67
tion. This process takes time to accomplish (meanwhile
that information is being spread around so that other
people are being induced to behave in a similar manner) .
Also the extent of borrowing would be limited and would
involve a great deal of time to convince people of the
desirability of the loan, to get funds over the value of
the portfolio the individual holds. Thus, it seems un
likely that any one individual could move the price of
an asset significantly even in the light of perfect
knowledge. Information must therefore be disseminated
to a significant number of people, or in the case of
holding companies, mutual funds, and so forth, a signif
icant number of controllers of a large number of people
before the price can reach a new equilibrium level.
b. The Problem of Imperfect Information:
As explained earlier, when an individual receives
a piece of information he makes two probability assess
ments: the first involving the reliability, and the
second relating to the associations that can be gleaned
from that information. Associated with these probability
assessments is a probability density function which can
be described by its mean (expected value) and variance.
Each individual is assumed to at least implicitly place
68
a value for the variance. That is to say, individuals
will attempt to estimate what the value of the informa
tion they received is by assessing the reliability,
potential associations and accuracy of those associa
tions in order to maximize their wealth. This process
can be described by a conditional probability statement,
i.e., the probability of an event having a particular
effect given that the information is correct, or
P(y/x) = P(y,x)/P(x) (5)
where x is the outcome (expected) based upon the informa
tion being correct (assuming that x and y are independent
and P(x) ^ 0).
The fact that information is not always perfect
is another reason that prices do not adjust immediately.
Take, for example, the case of an institutional arrange
ment where a piece of information is transmitted, say by
some form of mass media, to all individuals at the same
time. The fact that the piece of information was re
ceived at once does not necessarily mean that it will
change any market prices at all let alone change them
immediately to a new equilibrium. This may result in
cases where the expected value of the information being
correct and the expected event occurring is less than the
associated transactions costs in adjusting the portfolio
69
plus an amount to cover the risk involved. Until other
pieces of information are made available to either in
crease the expected value or decrease the risk premium
or both, the prices of the particular items the informa
tion may concern may not change at all. Any variations
in the expected value and/or variance from that of the
case of perfect information may move the price of the
(potentially) affected good (assuming that the expected
value is greater than the transactions cost and risk pre
mium) toward or away from but not necessarily to the full
information price. Only over a longer period of time
will prices be moved to an "equilibrium" as more and more
information is gathered about the market concerned, which
will adjust the expected values and risk levels to their
equilibrium levels.
The problem of price adjustments becomes even
more complex when different individuals with different
risk preferences, sources or information, and expecta
tions of outcomes are introduced. In the case of differ
ent risk preferences, those with lower aversions to risk
may be induced to adjust their portfolios with a smaller
expected value than those with higher risk aversions.
Because of these differences in risk preference, the
price will not adjust immediately to its "equilibrium"
70
level. It may follow a process such as the following:
As persons with lower risk aversion enter a market that
new market information or associations lead them to be
lieve a price change has or is expected to occur, the
price of the good begins to move toward a new equilib
rium. As this occurs,the expected value and/or risk to
everyone else will probably change because the price has
changed. Some amount of risk will be reduced if a price
movement is in an expected direction. (The price change
that has occurred when the people with lower risk aver
sion entered the market is another piece of information
that will have to be disseminated, and this will involve
additional time.) At a lower risk, say, the next set of
risk takers, a bit less adventurous than the first set,
enters the market pushing the price further toward its
new equilibrium. Again the price change leads others to
review the situation and after another information lag
(i.e., that the price has changed again), the next group
of risk takers enter the market. This process, which
may vary in length considerably depending upon how fast
people make decisions, the speed which information con
cerning price changes moves and the actual price changes
themselves, continues until all potential entrants of
all risk levels have had a chance to enter the market.
71
Other differences among consumers or consumer
groups such as differences in sources of information,
differences in expectations of outcomes, and the like,
can be handled in a similar manner as differences in risk
preferences. Again in every case where differences exist,
the entire price adjustment cannot be made immediately,
and the degree of differences among the consumers will
determine the degree of variations in the price adjustment
process.
The differences among consumers described above
have been shown to be a deterrent to the immediate price
adjustment process, even in the extreme case where in
formation is assumed to be provided immediately to every
one. The reason for this is that people will, in many
cases, not respond to a piece of information that may
affect their portfolio and wait until further information
is available and/or a definite price movement (seeing is
believing) before adjusting their portfolio. Since all
consumers do not have the same preferences for risk,
sources of information and expectations of outcomes, and
so forth, they will adjust their portfolios with varying
amounts of information or degree of price changes.
72
C. Spreading of Risk
The third influence to be included in the stand
ard analysis has already been explored and worked with
in many phases of economics and finance. However, it has
not explicitly been integrated into the analysis of
wealth distribution. This influence is the spreading of
risk.
It has been shown earlier (pages 35-41) that one
way of reducing the overall level of risk in the holding
of portfolio items is to diversify the assets and lia
bilities. From this earlier analysis, it is only a small
step to showing that those who have the largest port
folios are the ones who are able to diversify the most
and hence suffer smaller fluctuations in the value of
their portfolio (relatively) as a result.
This is because as more and more items are intro
duced, less relative risk for the entire portfolio is
experienced. Since risk reduction increases an indi
vidual's utility, those in higher wealth groups will
probably be better off when there are changes in the
price level and/or interest rates. This is because they
can better absorb these shocks by spreading them over a
variety of price and interest rate sensitivity port
folio items.
73
Summary
In order to gauge how government policy alters
the distribution of wealth in a static setting, two ques
tions must be answered. First, which assets and lia
bilities are affected (in what manner and to what degree)
by changes in the interest rate and the price level, and
second, who are the holders of these assets and liabili
ties (i.e., the distribution of portfolio composition).
Regarding the first question, many economists are
in agreement as to how, for example, inflation affects
different assets,^ or how changes in interest rates will
affect an individual's debt position. if the price
level should rise and/or interest rates should increase,
flows from wealth would be transmitted from net creditors
to net debtors.
Answers to the second question identify the
holders of the portfolio items affected by changes in
various market parameters. Nevertheless, either of these
questions do not yield much of a clue as to what happens
1;1-Kessel and Alchian, op. cit.
■^0. Brownlee, and A. Conrad, "Effects Upon the
Distribution of Income of a Tight Money Policy," Com
mission on Money and Credit. Stabilization Policies
(Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1963).
74
to the distribution of wealth when the element of time
(portfolio adjustment) is introduced. In order to gauge
policy under these circumstances, a third question con
cerning individual behavior must be asked, namely, why
do individuals desire to hold a particular bundle of
portfolio items at one specific time (or periods of
time) .
This last question is, of course, the one that is
of interest for this study. Traditional theory does not
attempt to explain why one individual or group may "wind
up" with those items that consistently lose value in
light of government policy while others may derive gains
in value over the same time period. It does not inquire
why different individuals hold different kinds of items,
what motivates them to make their particular portfolio
choices or what the effects of these choices will be on
the distribution of wealth (or vice versa) .
By exploring some of the parameters and decision
variables facing an individual when making (or planning)
a selection, the process of portfolio adjustment is
brought into sharper relief, and the most important fac
tors that determine whether an individual buys or sells
an item may be identified. From this kind of analysis
the effects of certain parameter changes such as those
75
induced by changes in government policy can be shown to
affect different individuals or groups in different
manners or to varying degrees and thus cause a redistri
bution of wealth among groups.
Four reasons were offered to explain why individ
uals in different wealth groups hold different assets and
liabilities and what the incentives were to switch their
portfolio composition. First of all, it was hypothesized
that those individuals in the highest wealth groups had
more information (i.e., had an incentive to acquire more
information) about markets relevant to their portfolio
than those in lower wealth groups. Because of this, they
would alter their portfolios earlier and easier (it
usually requires less effort to make a decision if more
information is available) than their counterparts with
less information, and would thereby avert losses and
enhance the value of their portfolio. Second, because
they have more information, the wealthier investor was
expected to make better predictions about the future
market performance of various assets and liabilities.
A greater accuracy of prediction will tend to enhance the
chances of increasing the value of his portfolio vis-a-vis
the less wealthy portfolio holder and also reduce his
risk in the future.
76
The third reason individuals in different wealth
groups hold different portfolio items is that those with
larger portfolios pay a lower per unit transaction price.
That is, the cost of moving in and out of a market would
be smaller (per unit) the larger the number of items
traded.
Finally, the fourth hypothesis deals with the
spreading of risk. It is held that the larger his
wealth, the greater the number of different kinds of
assets and liabilities an individual is induced to hold.
By holding many items, the potential loss (and gain) re
sulting from any one item is reduced to a fraction of the
less diversified portfolio.
It is a result of these differences among indi
viduals that changes in government policy may lead to a
redistribution of wealth. It would seem that those who
have the most information and who could anticipate best,
who pay the least relative price for transactions (port
folio adjustment), and who can spread their risks fur
thest are those with the greatest wealth. Hence, a
positive relationship between initial wealth holdings
and relative changes in wealth should be observed.
77
Finally, it would also seem that if these differ
ences among individuals did exist, their behavior pat
terns with respect to changes in interest rates and the
price level and portfolio investment decision making
should be significantly different.
CHAPTER IV
USING SURVEY DATA TO TEST FOR CONSUMER AWARE
NESS, ATTITUDES AND ANTICIPATIONS
The Data and Its Specifications
Employing data on the consumer sector collected
by the Survey Research Center of the University of Mich
igan, the first two hypotheses formulated in the previous
section dealing with wealth distribution resulting from
changes in various financial variables will be tested.
The data were collected from a panel study covering the
period 1967-1970 for a random sample of 1,649 American
families, where the same spending units were interviewed
for four consecutive years. (The problem of identifying
shifts in preference among individuals or groups of in
dividuals is alleviated because of the re-interview tech
nique and, hence, should provide more consistent cross
sectional results.)
Only the two years, 1968 and 1969, will be used
for the major portion of the analysis, however, because
they are representative of the kinds of government policy
conducted during the four-year period and because more
information pertinent to this study is available for
78
79
these two years (i.e., in terras of numbers of useful
questions asked).
Four types of questions asked of the interviewed
families are of importance for this study. They are
questions pertaining to: (1) stocks of wealth and income
from all sources, (2) information concerning the present
status of various market prices and the factors that
affect them, (3) the future values of the market prices
and the potential forces that may affect them, and (4)
selected consumer behavior patterns. A list of the ques
tions employed can be found in the question sets at the
end of this chapter.
From the information contained in the first set
of questions (Question Set A), each individual's (non
human) wealth will be calculated then the portfolios
will be ordered on the basis of net worth (assets and
liabilities) . Human wealth or human capital will not be
calculated because short-term changes in most financial
variables, especially the interest rate, have not been
shown to be significant influences upon human capital.1
^It should be noted that while the data on indi
vidual wealth components used in this study are the best
available, they are not a perfect measure of the value
of every item of an individual's portfolio because set
ting market values on portfolio items is difficult
assessment for many individuals. Nevertheless, a good
approximation of the total portfolio valuation can be
expected since the large sample in each decile (144)
80
After the wealth holders are ordered on the basis
of their initial wealth holdings (1968, 1969), they will
be divided into deciles to avoid the problem of setting
arbitrary limits for categories. From this division, a
mean valuation of all assets and liabilities in that par
ticular group will be found and will constitute the
"representative man's" portfolio for that group. A rep
resentative man will be found for each wealth group.
Testing the Hypotheses
Information and Antici
pations Hypotheses
The first hypothesis holds that there are differ
ent degrees and kinds of information that members of
different wealth groups possess. To test this hypothesis,
a chi-square test will be performed to determine if a
significant difference exists among wealth groups with
respect to their responses to various questions concern
ing current market activities. If significant differ
ences are found at the 1 percent level among wealth
groups, then this would indicate that wealth and answers
to a particular question are not independent.
makes the differences in total portfolio values less of
a problem when groups of individuals are compared and the
measurement errors are spread over a large base.
81
The amount and variety of information held by
individuals in different wealth groups will be manifest
by their responses to a set of selected information ques
tions (a list of these questions can be found in Question
Set B replete with a breakdown of responses). A chi-
square test will be made for each question selected.
The chi-square test is used to determine if the
various information questions are associated with the
levels of wealth possessed by the respondents. However,
it would be useful to know where the differences between
the wealth groups that make up the significant chi-
squares lie since these tests do not imply anything about
the direction or degree of association between wealth
and responses. In order to do this further analysis must
be undertaken. The first technique considered was an
analysis of variance with a least significant differences
test (based upon Scheffe "Contrasts" between means).
Unfortunately, this could not be employed because the
data are nominal. (Theoretically, the data could be con
verted by assigning arbitrary numbers to the answers,
e.g., YES =1, NO = 2, and so forth. However, the num
bers assigned could influence the results.) This left
the analysis of the contingency table known as parti
tioning or collapsing the table.
82
A contingency table can be partitioned into
R(rows) by C(columns) factoral different tables, so
a priori specifications must be made. Since wealth
groups' answers to questions are of importance, the per
centage of particular answers to a question should
hypothetically vary directly (correct answers) or in
directly (incorrect answers) with the mean wealth of
the group.
One approach to handling this problem is to work
with the columns (particular answers to questions) rather
than the rows (wealth groups) since the wealth groups
are forced row averages. This can be done by finding
the mean wealth of all individuals who answered a ques
tion one way then comparing that mean with the mean
wealth of those who answered the question another way.
Using a student's "t" test, significant differences be
tween the mean wealth of different respondents can be
determined. If the mean wealth of those who answered a
question correctly is significantly different at the 1
percent level from those who answered it incorrectly,
tested by the student's "t" test, the direction of the
association between wealth and answers to questions can
be suggested (i.e., in this case, the association of
correct answers and wealth, as hypothesized, would be
positive).
83
The Probability Corre
lation Method
The second method used to analyze the contingency
tables involved correlating the mean wealth of a group
with the probability of choosing a particular answer.
The associated probability is derived from the relative
frequency of the answer to all answers for a particular
question. For example, if one wealth group had 60 "yes"
answers, 100 "no" responses, and 40 "maybe" answers to a
question, the probability of that group's answering yes
would be 30 percent, no 50 percent, and 20 percent maybe.
The relative proportions of responses for each wealth
decile are found for each answer, and they are corre
lated with the corresponding wealth groups. Strong and
positive correlations between the mean wealth of these
groups and the proportion of total responses to an an
swer would indicate a strong association between the
answer and wealth, and vice versa for strong negative
correlations. Since there are ten deciles, there must
necessarily be ten observations per answer (proportional
responses) leaving only eight degrees of freedom. Hence,
the significance levels of the correlations may not be
very high. Notwithstanding these small sample problems
the correlation coefficient should possess the correct
84
sign and the significance level should be higher the
stronger the correlation with wealth (positive or nega
tive) .
The employment of this method allows for a row
analysis of the contingency tables whereas the previous
mean wealth analysis focused on the columns. Both tech
niques are designed to appraise the relationship between
wealth and specific answers to questions. In the corre
lation analysis, it is expected that correct answers to
information and anticipation questions will have the
largest positive coefficients of correlation, i.e., be
explained best by wealth, and vice versa for incorrect
answers.
The correlations with the wealth deciles dis
cussed above are based upon relative frequency distri
butions of particular answers.2 They are not arbitrary
assigned values to the answers which have an influence
on the test.^
2See A. E. Maxwell, Analyzing Qualitative Data
(New York: John Wiley & Sons, Inc., 1961), Chapter 4."
^In a linear regression model such as
Yi = aX^ + e^, setting the dependent variable (i.e., the
answer) equal to zero for yes and one for no would force
the error term (ei) to be correlated with X (the inde
pendent variable, wealth).
85
The question set selected here for testing the
first hypothesis was chosen on the basis of the kind of
information relating to current market phenomena the
questions inquire about. No specific training is re
quired to answer these questions so that individuals in
all wealth categories had a chance of answering it cor
rectly. The questions' ability to encompass both broad
and specific areas should make them a useful barometer
of individuals' information.
It should be noted that the purpose(s) for
which these questions were asked (in the consumer survey)
are not the same as those set but for this research.
Hence, in some cases the breadth is sometimes too wide
and the scope sometimes too narrow. Also these questions
are but a small fraction of the large number of questions
(400-500 questions) asked during a long and extensive
interview. The placement of the questions in the inter
view (usually either in the beginning or at the end) may
be another factor that might affect the accuracy of the
answers. Nevertheless, these questions will be used in
this analysis because they are the best available source
of data for gauging individuals' knowledge of market's
conditions and activities.
86
The first information question deals with busi
ness conditions in general during the current (survey)
year. In order to determine which response came closest
to a "correct" answer for a particular question, outside
sources will be consulted. In this case the sign change
of the Federal Reserve Board's Business Index will be
employed as a gauge of whether business conditions im
proved, stayed the same or became worse. The change will
be calculated by subtracting the value of the index in
the year preceding the survey from the value in the
current year.
The second question relates indirectly to a knowl
edge of the market interest rate and specifically to the
borrowing rate on durable goods. Answers to this ques
tion may vary greatly for two reasons. First, there are
many sources of borrowing for durable goods (even at the
same institution depending upon the kind of product pur
chased, e.g., a new vs. a used car, the percent down
payment, if collateral is required, and so forth). Sec
ond, individuals may not know how much they really pay
for a loan (the truth in lending program has been set
up to reduce this kind of ignorance).
Another question that relates to the general
knowledge of market interest rates is number three, but
87
here the rate is on income yielding assets (on the lend
er's side). An average of saving account and certificate
of deposit rates will be used as the market savings rate.
The fourth question should indicate an awareness
of changes in interest rates in the market and will serve
as an indirect check on the other interest rate questions
(for errors that may have occurred due to misspecifica-
tion or misinterpretation). Both years in question had
significant changes in interest rates (measured by both
prime rates and government bond rates).
The next four questions are derived from one
rather broad question relating to favorable or unfavorable
business conditions (see question number 5). The answer
to the first part of this question is of minor importance
to this study but how they came to that answer (i.e.,
the second part of the question) may be very useful. The
four derived questions ask, in effect, what has happened
recently to employment (responses 1, 5), prices (re
sponses 2a, 2b, 6a, 6b), profits (responses 3a, 3b, 7a,
7b), and the stock market (4, 8). The answers to these
questions will be checked against Department of Commerce
figures for unemployment, the consumer price index for-
prices, Department of Commerce reported corporate profits
for profits and the Standard and Poor's Stock Index for
88
the stock market to see which responses corresponded
closest to what actually took place.
The second hypothesis will he tested in a similar
fashion to the first one and the answers to the questions
treated in a like manner. A second set of questions will
be used to test this hypothesis (see Question Set C)
which should yield an indication of how well individuals
anticipate selected market activities.
The first question deals with the general eco
nomic picture a year from the time of the interview. To
determine how well the future was predicted real gross
national product (GNP) will be used because this question
requires an answer that reflects total wages, profits,
rents and other incomes. Whether to use real (deflated)
or nominal GNP may ostensibly present a problem because
of the word "financially" used in the question (in 1968
it makes no difference, but in 1969 real income declined
while nominal income rose). Real GNP is chosen because
"financial" does not necessarily connote merely take home
wages, and so forth, i.e., people are believed to look at
price changes as a deflator of financial worth, especial
ly after periods of ensuing inflation (which occurred
prior to the years in question).
89
Questions two and three focus on one important
financial variable, namely prices, and should give a
good indication of how well different groups and indi
viduals are able to predict future price levels. For
a measure of price level changes, the consumer index
(1958 base) will be used. This gauge was selected be
cause consumers are more aware of prices that they see
everyday in the supermarkets, and so forth, than say
steel prices which are used in the calculation of whole
sale prices and because information about price changes
(in newspapers, television, and so forth) are usually
in terms of consumer products.
The next question is concerned with business con
ditions and how well various individuals are able to
foresee their outcomes. The answers will be compared
with the Federal Reserve Board's Index of Industrial
Production.
The fifth question is somewhat related to the
previous one concerning business sectors of the economy
differently and different industries and income groups
in a variety of ways that may vary inversely with busi
ness conditions (especially where lags are introduced) .
Department of Commerce unemployment for civilians figures
will be used to verify the answers to this question.
90
The final question will be used to gauge the
facility of prediction of one of the most important mone
tary policy variables, namely the interest rate, which
may have a strong effect on the portfolio valuation of
any wealth group. They will be measured by the prime
rate twelve months after the interview since consumers
are, generally, more aware of borrowing rates rather
than lending rates.
Transactions Costs
and Risk Spreading
The third reason for individuals in different
wealth groups holding different portfolio items was that
those in higher wealth groups faced lower transactions
costs (per unit of item traded). This reason was not
tested because of the paucity of situations where the
alternative situation even partially existed. The anal
ysis in the previous chapter suggests that earlier
transactions are cheaper per unit of item traded in al
most every case. (The exception progressive marginal
transactions costs being found only in the rarest cases.)
Since those with the largest portfolios can make the
largest adjustments, their per unit transaction cost is
thus lowest.
91
It follows that if those with the largest port
folios make the largest transactions then those in the
higher wealth groups can move relatively more easily
into and out of portfolio items than those in lower
wealth groups. This greater ease of movement (less
costly) may afford certain individuals the ability to
purchase or sell items that others could not afford due
to the higher transactions cost.
Finally, the last hypothesized reason for differ
ence in portfolio composition among wealth groups was
the element of risk spreading. This hypothesis will
also not be tested here because of the breakdown of
assets is not sufficient to estimate the variety of
assets held by individuals. However, the relationship
between large wealth holders and greater diversification
has already been established.4 Hence, it would not be
unusual to find not only different kinds of items among
wealth groups but also larger numbers of them the greater
the amount of the individual's wealth.
^See, for example, James Tobin, “Liquidity Pref
erence as Behavior Toward Risk," Review of Economic
Studies (February 1958); Basil T. Moore, An Introduction
to the Theory of Finance (New York: Free Press, 1968);
John Litner, "Security Prices, Risk and Maximum Gains
from Diversification," Journal of Finance (December 1965);
Harry M. Markowitz, Portfolio Selection; Efficient Diver
sification of Investments (New York; John Wiley & Sons,
Inc., 1959).
92
Wealth Maximization
Hypotheses
The first two hypotheses relate the wealth of the
consumer to the amount of information he possesses con
cerning market phenomena potentially affected by changes
in interest rates or the price level and his ability to
anticipate the future direction of these variables.
Significant differences among wealth groups would indi
cate that those individuals in higher wealth brackets
purchase more information and are more aware of levels
and changes in important financial variables as well as
their causes and their potential effects (as evidenced
by the ability to predict better the future values of
important financial variables).
In order to see if, in fact, those with greater
information did do better with their portfolios (doing
better being defined as increased returns, lower risk or
both) than those in lower wealth groups, the following
hypotheses should be tested.
First, individuals in higher wealth groups are
expected to realize larger positive changes (smaller
negative changes) in their portfolio value than those in
lower wealth groups.5 Hence, the relation between
^By using wealth groups all four of the above
hypothesized effects (i.e., information, anticipations
93
initial wealth levels and relative changes in wealth in
each wealth group should be found to be positive and
significant. This is not the whole story, however. The
other component in the maximization of the utility of an
individual's portfolio is the minimizing of risk. Un
fortunately, there is not any way that this can be tested
using cross-sectional data (it is possible using time
series data but this is not available), because there is
no measure of risk preference or avoidance vis-a-vis
returns for an individual or group. Due to the lacuna in
the data, it is difficult to tell how wealth (as measured
by utility) is being redistributed. The reason for this
is that risk and return are traded off (i.e., substitutes)
and hence both must be examined together in order to tell
if those with the greatest amount of information, and so
forth, did increase their utility (increase their re
turns and/or reduce their risk) more than those with less
information, and so forth.
Portfolio Adjust
ment Hypotheses
So far the hypotheses tested have dealt with the
amount of information individuals possess and how well
as well as transactions costs and risk spreading) would
be included in such a test.
94
they can anticipate future market outcomes. This implies
that individuals in different wealth groups manifest dis
similar patterns of behavior. It would be interesting to
see if the survey data could corroborate this implication.
Four questions relating to consumer behavior with respect
to portfolio adjustments will be explored (see Question
Set D). It is hypothesized that the behavior (i.e., re
sponses to these questions) of different wealth cate
gories) will not be the same. This will be tested in a
similar fashion as the information and anticipations
hypotheses in the first section.
Affirmative responses to the first question indi
cate an awareness of the effects of interest rates upon
the value of the portfolio, and, moreover, an overt will
ingness to do something about it to offset losses or
realize gains by shifting portfolio composition. There
may be some ambiguity in using this question. The last
phrase "because of changes in interest rates" may be
taken to mean actual and/or expected changes. For pur
poses of this study both interpretations are desirable.
There may be cases where individuals changed their port
folio because of expected changes in interest rates
which would indicate an awareness of the effects of in
terest rates (perhaps an even better understanding of the
95
market forces) but they might answer in .the negative.
Hopefully, these cases are few, but there is no assurance
that they are. As a consequence of this alleged ambi
guity (if it is significant), the results may be biased.
The second question is more specific in deter
mining the reaction to changes in interest rates as they
relate to the particular action taken as a result of
these changes.
The third question involves a hypothetical situa
tion where an individual is asked to make an investment
decision at one particular time given what he knows about
current market activities and his expectations.
Finally, the reasons behind this investment
decision are explored in the fourth question.
96
QUESTION SET A
NET WORTH 1968
Assets
1. Present value or cost of house.
2. What is the total amount you now have in all these
(checking) accounts?
3. Savings accounts net of certificates of deposit.
4. What is their (certificates of deposit) total
value?
5. Approximately how much are these stock and shares
worth?
6. About what is the face value of your bonds alto
gether?
7. Do you own any real estate such as a lot, summer
home, an apartment building, or business property
(other than this place here)?
Liabilities
1. (6-digit Ml Box) Code first mortgage Ml Box
(dollars)
2. (6-digit M2 Box) Code second mortage M2 Box
(dollars)
3. Remaining Installment Debt incurred in 1967 on
all A & R
4. Remaining Installment Debt incurred NOT in 1967
on A & R
5. Remaining Installment Debt incurred in 1967 on
ALL cars
6. Remaining Installment Debt incurred NOT in 1967
on ALL cars
97
7. Remaining Installment Debt incurred in 1967 on
ALL durables
8. Remaining Installment Debt incurred NOT in 1967
on ALL durables
9. Remaining "Other" Installment Debt incurred in 1967
10. Remaining "Other" Installment Debt incurred NOT
in 1967
11. Remaining Medical and Dental Installment incurred
in 1967
12. Remaining Medical and Dental Installment incurred
NOT in 1967
13. Remaining Non-Instailment Debt on A & R (all years,
all A & R)
14. Remaining Non-Installment Debt on cars (all years,
all cars)
15. Remaining Non-Installment Debt on durables (all
years, all durables)
16. Remaining Non-Instailment Debt on "other" (all
years)
17. Remaining Non-Installment Debt on medical and
dental (all years)
18. Amount borrowed on real estate bought during last
twelve months
19. Do you owe any money on the property which you
bought before 1967? How much do you still owe
(not counting debt on property purchased in 1967)?
20. Amount owed on stock
Nonhuman Income
1. Mixed labor capital income
2. Total capital income
98
NET WORTH 1969
Assets
1. Present value or cost of house
2. About what is the total amount you now have in all
these (checking) accounts?
3. About what is the total amount you have in all
these (savings) accounts?
4. What is (your certificates of deposits') total
value?
5. Approximately how much are these stock and shares
worth?
6. About what is the face value of your bonds alto
gether?
7. Do you own any real estate such as a lot, summer
home, an apartment building, or business property
(other than this place here)?
Liabilities
1. Amount of unpaid balance on first mortgage
2. Unpaid balance on second mortgage
3. Remaining Installment Debt incurred in 1968 on
all A & R
4. Remaining Installment Debt incurred NOT in 1968
on A & R
5. Remaining Installment Debt incurred in 1968 on
ALL cars
6. Remaining Installment Debt incurred NOT in 1968
on ALL cars
7. Remaining Installment Debt incurred in 1968
on ALL durables
99
8. Remaining Installment Debt incurred NOT in 1968
on ALL durables
9. Remaining "Other" Installment Debt incurred in 1968
10. Remaining "Other" Installment Debt incurred NOT
in 1968
11. Remaining Medical and Dental Installment Debt
incurred in 1968
12. Remaining Medical and Dental Installment Debt
incurred NOT in 1968
13. Remaining Non-Instailment Debt on A & R (all years,
A & R)
14. Remaining Non-Installment Debt on cars (all years,
all cars)
15. Remaining Non-Installment Debt on durables (all
years, all durables)
16. Remaining Non-Installment Debt on "Other" (all
years)
17. Remaining Non-Installment Debt on Medical and
Dental (all years)
18. Amount owed on stocks
19. Amount borrowed on real estate bought in last
twelve months
20. Do you owe any money on the property which you
bought before 1968? How much do you still owe?
Nonhuman Income
1. Mixed labor capital income
2. Total capital income
100
QUESTION SET B
INFORMATION QUESTIONS
1. Would you say that at the present time business con
ditions are better or worse than they were a year aero?
1. Better now (41%)
3. About the same (40%)
5. Worse now (17%)
8. D.K. , depends (1.7%)
9. N.A. (.4%)
2.A.Suppose you needed a thousand dollars for a car which
you would repay in twelve monthly payments, about how
much do you think the interest or carrying charges
would be?
NOTEs If respondent mentions more than one source and
mentions different rates for the sources, code
the lowest rate. Thus "10 percent from the car
dealer and 7 percent from the bank would be
coded 3. (6.6-9.5%)."
Code Per Year Interest (interest rate on $1,000)
1. Under 4% or $39 or less per year (2%)
2. 4 - 6.5% or $40 - 65 per year (25%)
3. 6.6 - 9.5% or $66 - 95 per year (28%)
4. 9.6 - 12.5% or $96 - 125 per year (14%)
5. 12.6 - 15.5% or $126 - 155 per year (3%)
6. 15.6 - 20.5% or $156 - 205 per year (6%)
7. 20.6% or over or $206 or over per year (4%)
8. D.K. (6%)
9. N.A. (3%)
(If both dollar figure and percent given, code
percent.)
B.Can you give me a rough estimate of what you think
the charges might be?
(See Note above.)
Code per Year Interest (interest rate on $1,000)
1. Under 4% or $39 or less per year
2. 4 - 6.5% or $40 - 65 per year
101
3. 6.6 — 9.5% or §66 - 95 per year
4. 9.6 - 12.5% or §96 - 125 per year
5. 12.6 - 15.5% or §126 — 155 per year
6. 15.6 - 20.5% or §156 - 205 per year
7. 20.6% or over or §206 or over per year
8. D.K.
9. N.A.
0. Inap.
(If both dollar figure and percent given, code
percent.)
3. What is the (highest) rate of interest you earn on
your savings account(s) (or certificates of deposit)?
XX. Actual annual rate of interest in tenths of a
percent (e.g., "5^" - 55)
97. 9.7 percent or more
98. D.K.
99. N.A.
00. Inap. owns no savings accounts (coded 5 or 9
in col. 15)
4. Do you happen to know whether there have been any
changes during the last few months in the interest
rate paid on savings, or in the interest paid by
individuals or businesses when they borrow money?
(If Yes) What kinds of changes?
1. Increase (69%)
3. Some increases, some decreases (1%)
5. Decrease (2%)
6. Mentions change, D.K., N.A. direction (7%)
7. No, no change; interest rates have not changed
(or changed very little and nothing codable
above); "No, I haven't heard of any." (9%)
8. D.K. whether rates have changed; "No" (19%)
9. N.A. (1%)
5. During the last few months, have you heard of any
favorable or unfavorable changes in business condi
tions? (If Yes) What did you hear?
Favorable Changes
1. Employment has risen, is rising; more overtime;
plenty of jobs or work around; unemployment
declining
102
2a. Lower prices; prices aren't rising; stable
prices
2b. Higher prices; inflation; prices rising
3a. Profits high, higher
3b. Profits low, lower
4. Stock market
Unfavorable Changes
5. Drop in employment; high or higher unemployment;
layoffs; less overtime; people are working
short hours
6a. Prices are falling, will fall, are too low;
deflation
6b. Prices are too high, are going up; inflation
7a. Profits high; too high
7b. Profits low, falling; businesses losing money,
failing, going bankrupt
8. Stock market; decline in stock prices
103
QUESTION SET C
ANTICIPATIONS QUESTIONS
1. Now looking ahead — Do you think a year from now,
you people will be better off financially, or worse
off, or just about the same as now?
1. Better (44%)
3. Same (43%)
5. Worse (5%)
8. Uncertain, D.K. (7%)
9. N.A. (1%)
2. Talking about prices in general, I mean the prices
of the things you buy — do you think they will go
up in the next year or so, or go down, or stay
where they are now?
1. Will go up (85%)
3. Stay the same (13%)
5. Will go down (2%)
8. D.K. ; uncertain, depends (1%)
9. N. A. ; R speaks only of hopes or wishes
3. IP WILL GO UP — How large a price increase do you
expect? Of course nobody can know for sure, but
would you say that a year from now prices will be
about 1 or 2% higher, or 5%, or closer to 10% higher
than now, or what?
1. 1% (13%)
2. 2% (19%)
3. 3% (11%)
4. 4% (3%)
5. 5% ( 26%)
6. 6 to 9% (2%)
7. 10% or more (7%)
8. D.K. (3%)
9. N.A.; "Some"; "A little"; "A lot" (1%)
0. Inap. (14%)
104
4, And how about a year from now, do you expect that in
the country as a whole business conditions will be
better or worse than they are at present. or just
about the same?
1968 1969
1. Better a year from now 24% 22%
3. About the same 58 63
5. Worse a year from now 11 12
8.
9.
D.K.; depends
N.A. (absolute answers "good"
5 4
or "bad") 2 1
5. How about people out of work during the coming twelve
months — do you think that there will be more unem
ployment than now, about the same, or less?
1. More (27%)
3. About the same (55%)
5. Less (15%)
8. Don' t know (2%)
9. N.A. (1%)
6. No one can say for sure, but what do you think will
happen to interest rates during the next 12 months?
1. Go up (41%)
3. Stay the same (40%)
5. Go down (5%)
8. Don't know (11%)
9. N.A. (1%)
105
QUESTION SET D
BEHAVIORAL QUESTIONS
1. Considering your financial assets, since this time
last year did you do anything because of changes
in interest rates?
1. Yes; NA what R did (1%)
2. Switched from one place to another; bought and
sold (3%)
3. Bought (1%)
4. Sold (1%)
5. No (50%)
8. Don11 know
9. Not available (owns more than $1,000 worth of
assets) (2%)
0. Inap. (has less than $1,000 in savings) (42%)
2. What did you do? (with regard to changing assets in
response to interest rate changes)
Code 2 Mentions
1. Bought stocks (1%)
2. Sold stocks (1%)
3. Bought bonds (1%)
4. Sold bonds (1%)
5. Withdrew from savings account (1%)
6. Added to savings account (1%)
7. Reduced checking account
8. Bought certificates of deposit (1%)
9. D.K. , N.A. (owns more than $1,000 worth of
financial assets) (1%)
0. Other; Inap. (has less than $1,000 worth of
financial assets) (93%) *
106
3. Suppose you had some new savings. What would be the
wisest thing to do with the money — put it in a
checking account or savings account, buy a govern
ment savings bond, invest in real estate, buy common
stock, or what?
1. Checking account or savings account (including
certificate of deposit) include certificates
of deposit, credit unions, Christmas clubs
(43%)
2. Buy bonds (including government savings bond)
(13%)
3. Invest in real estate (23%)
4. Buy common stock or mutual fund shares,
preferred stock, investment trust (17%)
5. Other (include repayment of debts) include
"pay bills" (3%)
8. D.K. , N.A. )
0. Inap., no second mention )
4. Why do you make that choice? (major reasons for
first choice)
01. Safety; no chance of loss (25%)
02. (High) rate of return, (high) interest,
(high) yield (27%)
03. Liquidity, easy to get in case of emergencies
or other needs, convenient, accessible (11%)
04. Capital gains, appreciation of investment (11%)
05. Hedge against inflation (3%)
06. Make more money, NA whether codes 2, 4 (5%)
07. Other (2%)
08. D.K. (2%)
09. N.A. (12%)
00. Inap., no second mention (1%)
CHAPTER V
RESULTS OF THE TESTS
In this chapter, the results of the empirical
tests will be presented and the more important implica
tions and relationships will be drawn from the data.
For the most part, the testing done here was re
lated to the "representative man" or mean valuation of
wealth in each decile. The following Table 1 summarizes
the mean portfolio valuation in each decile that will be
referred to in the remainder of this chapter
TABLE 1
PORTFOLIO VALUES OF REPRESENTATIVE MEN
YEARS 1968, 1969
Decile
1968
Mean Wealth
1969
Mean Wealth
1 -2,956 -3,280
2 -266 -188
3 611 1,108
4 2,634 3,657
5 5,229 6,749
6 8,745 10,532
7 13,085 14,986
8 18,608 20,964
9 31,485 35,060
10. 104,438 106,231
107
108
All analyses conducted with questions in 1968 are
tested with the mean decile for that year and likewise
for 1969. This is done because of the differences be
tween the portfolio valuation of the representative men
in these years.
Tests of the Hypotheses
As outlined in the previous two chapters, access
to information and the ability to predict market activ
ities are hypothesized to be related to wealth in a
direct and positive manner. The main tool for this anal
ysis was the chi-square test of wealth deciles with an
swers to various questions relating to market activities
both past (information access) and future (ability to
predict) .
The Chi-Square Tests
A chi-square test is used to test for associa
tions between two methods of classification of the
members of a population. In this case, the population is
a random sample of 1,436 consumers in America, and the
classifications are Wealth Groups and Answers to Various
Questions (see Question Sets B, C, and D) .
Answers to twenty-three questions were compared
with the appropriate representative men. These include
109
questions concerning interest rates and the price level
as well as employment, business conditions, national
productivity, and other such questions. The reason the
latter set of questions were included in the tests was
to compare the influence of wealth on answers to inter
est rate and price level questions with questions re
lating more to other Kinds of information that may influ
ence an individual's wealth. The following are the
results of the chi-square tests. (The percentage break
down of answers can be found in parentheses next to the
respective answer.)
For hypothesis one, the questions in Question
Set B were tested. Question number one, concerning
present business conditions and their relationship to
the previous year's condition for both 1968 and 1969,
were found to be insignificant. In 1968 about 80 percent
answered better or the same, and in 1969, 86 percent
answered better or the same. Two national indicators,
GNP and the FRB index of industrial production both
moved up during these periods corroborating the answers.
This would indicate a high level of awareness on the
part of consumers concerning how American business is
doing. It would also imply that this kind of information
is made available cheaply enough for the vast majority
to afford it.
110
The second question (2A) tested concerned the
interest rate on borrowing $1,000. Here the chi-square
statistic was significant at the 1 percent level. This
would imply that wealth and borrowing rates on loans are
associated in some non-random fashion.
Question 2B was not used even though the chi-
square statistic was significant because of the large
number of total responses taken up by 2A.
The third question, what is the highest rate of
interest you earn on savings or C.D.s, was also found to
be significant at the 1 percent level. The largest group
(35 percent) had to be excluded because they had no
savings. Prior to April, 1968, Regulation Q did not
allow interest rates over 5^ percent to be paid by mem
ber banks irrespective of size or maturity. since then,
up to 6% percent is allowed on large deposits for over a
half a year. The usual rate at commercial banks was
still 4 - 5 percent throughout 1968-1969, while Certifi
cates of Deposit and saving and loan associations paid
slightly higher rates. The percent distribution of re
sponses seems to fit closely with the market interest
rates indicating a high level of awareness of market
interest rates by savers.
Ill
The next question which involves changes in in
terest rates also was found to be significant at the 1
percent level. A very large 68 percent were aware of
the interest rate increase, indicating an awareness of
interest rate trends. However, 19 percent said they did
not know about any change in the rate. These strong dif
ferences in responses may be explained partially by the
wealth differences of the respondents.
As for the last question, the breakdown of four
questions from one did not work well and all results were
insignificant. This was probably due to the large number
of alternative answers to the questions.
The second hypothesis was tested by using the
questions in Question Set C. The first question was
found to be insignificant for both 1968 and 1969, as
approximately 86 percent of the interviewees responded
that financial conditions would be better in the coming
(rather than current) year.
What is interesting to note is that in 1968-69
GNP (the indicator used to measure if people are better
off) went up but in 1969-70, it went down slightly in
real terms. Hence, the 44 percent that said "better"
in 1969 were not correct while the 46 percent saying
"better" in 1968 were. This indicates that even though
112
a large percentage of responses were correct in one year
the importance of new information and revised anticipa
tions cannot be ignored. The fact that new information
and revised anticipations take time to be disseminated
and formulated when changes in the system are encountered
helps explain the answer patterns in these years.
The next two questions relate to price level
changes (predicted). The first is found to be insignif
icant with respect to wealth as most individuals (84
percent) expected the price level to increase. On the
other hand, just how much inflation would be encountered
was found to differ significantly by wealth groups.
People's ability to predict the rate of inflation
(anticipated) is shown here to also be associated with
their initial wealth holdings.
The next question concerning projected business
conditions is insignificant for both 1968 and 1969.
Approximately 23 percent said "better" in both years,
where they were correct for 1968 but incorrect for 1969
(measured against the FRB index). These results are
similar to the first question's and would also lend
support to the implication that following a trend (i.e.,
last year's results) is not a good predictive tool and
hence, new information and the ability to readjust an
ticipations are important factors in projecting the
113
future outcomes of market activity.
The question concerned with predicting more or
less unemployment was also found to be insignificant.
The insignificant chi-square statistic indicates that
there was little association between wealth groups and
answers to this question. This implies that people watch
unemployment closely (information is available cheaply
enough) in all wealth groups and no one has an advantage
in predicting it because of more or less wealth.
Finally, an interest rate projection is called
for in the last question. It is significant at the 1
percent level. Interestingly enough, 92 percent either
did not know or guessed wrong. Nevertheless, wealth was
a significant factor in the respondents indices.
Reviewing both sets of questions, it seems that
an individual’s ability to predict and knowledge of
financial variables (interest rates and the price level)
seem to be significantly related to wealth, while real
variables (employment, etc.) were not or only to a small
degree related to wealth. This would imply that the
financial variables' changes have strong effects on
wealth, as hypothesized. Before discussing what the re
lationships between wealth and these questions are, an
other set of questions concerning actual and potential
consumer behavior will be considered in light of the
114
hypothesis that individuals in different wealth groups
behave differently with respect to government policy.
Questions pertaining to the hypothesized behav
ioral differences among wealth groups can be found in
Question Set D. The first one inquires about changes
in the portfolio composition due to changes in interest
rates. The answers were significantly associated with
wealth levels (at the 1 percent level), even though 42
percent had less than $1,000 in savings and hence were
excluded. Only 5.5 percent responded that they did ad
just their portfolio because of interest rate changes
while 50 percent said they did nothing about it. Even
though there was only a small percentage answering yes
the significant chi-square statistic would lead one to
believe that the differences among wealth groups of
these respondents was great which in fact does show up
in the chi-square analyses below.
The second part (the second choice) of this ques
tion had to be dropped because of the large number of
inappropriate responses (about 96 percent was taken up
by the first part of the question).
The second question (both parts) was also dropped
due to a large number of inappropriate responses.
115
The third question asked what the wisest thing
to do with new savings was. Here, the chi-square was
significant at the 1 percent level. The varied responses
thus revealed an association between wealth and what in
dividuals would do with additional (marginal) investment
resources. This association could be due to differences
in information and/or anticipations among wealth groups.
The second response to this question was dropped because
of the paucity of responses.
The last question inquired into the motives for
making the investment choice in the previous question.
Here the chi-square statistic was also significant at
the 1 percent level implying that wealth and motivation
for investment were related. The distributional break
down of the question fell primarily into safety (25 per
cent) and returns (27 percent) with 11 percent looking
for liquidity and 11 percent seeking capital gains.
Hence, about half of the respondents looked for increased
return while the other half for risk reduction.
Analysis of the Chi-Square Tables
The first technique used to analyze the contin
gency tables was the mean wealth test. Using a student's
"t" test, significant differences in the mean wealth of
116
respondents to a question could be determined. Moreover,
the direction of the association between correct answers
(where applicable) and wealth could also be detected.
Only questions that had significant chi-squares were
scrutinized in this fashion.
In the first set of questions (information),
three signficant sets of responses (questions 2, 3, and
4) were found.
The mean wealth of those responding 6-9 percent
as the interest costs for borrowing a $1,000 was signif
icantly greater than those responding 4-6 percent
($24,774 vs. $17,563). This would ostensibly imply that
those in the higher wealth groups paid more for a loan
than those in lower wealth groups. However, a more
plausible reason is that those in lower wealth groups
were not as aware of the real cost of the loan, i.e.,
the wealthier were more aware of the real cost of the
loan. Simple interest rates usually vary between 5
percent and 8 percent for collateral loans whereas they
range from 10 percent to 18 percent on compounded loans
(not backed by collateral). Dealers may state that the
cost of a loan is 4 to 6 percent but this is not the
entire interest cost.
The next "t" test showed a greater (but not sig
nificant) mean for those responding 6-9 percent than for
117
those saying 9.6-12.5 percent ($24,774 vs. $21,554) im
plying that those in higher wealth groups would pay a
lower rate on a $1,000 loan. Two observations can be
made about these results: (1) individuals in higher
wealth groups are more aware of where and how to get
lower cost loans; or, (2) that differences in risk
(assessed by the lender) account for this.
The second significant question further pursued
the individual1s knowledge of market interest rates on
the lending side. The mean wealth of those answering
5-6 percent return on savings was $49,609. This was
significantly greater than those answering 4-5 percent
($23,808) and those answering 3-4 percent ($18,132) .
While these last two groups were not significantly dif
ferent (i.e., 3-4 percent and 4-5 percent) the group
responding with the higher return had the higher mean
wealth. This implies that those with the highest amount
of wealth are those who get the highest returns on their
savings. This may be a result of having more market
information, e.g., concerning C.D.s, or because of the
larger amounts of savings which will get higher returns.
The second reason also points to differences in
information levels among wealth groups because banks will
pay higher returns on larger deposits to keep them from
118
switching their accounts to more lucrative savings in
stitutions. If this is so, the individuals in higher
wealth groups should be more willing to move as a result
of interest rate differences. This is supported in the
breakdown of question 2 of the behavioral set (below) .
Hence, it is the knowledge of various market interest
rates that determines the return on savings and this
seems to be positively related to wealth.
The other significant question refers to changes
in the market rate of interest. The mean wealth of those
answering correctly was significantly greater than those
who did not know about any changes. It was greater (but
not significantly) than those who answered no change
($23,809 vs. $13,975) and also larger than those who said
it went down ($23,809 vs. $11,471, but also insignifi
cant) . Hence, it seems that knowledge of interest rates
changes is directly related to wealth. Even though the
significance tests do not show a difference at the 1
percent level, the means are very different (60-100 per
cent) and greater the closer to the correct answer. The
reason for the large significance level is due to the
diversity of the two samples in terms of size.
On the basis of the chi-square analyses performed,
there were no cases found which contradict the first
119
hypothesis and many cases where it is supported.
The second set of questions relating to antic
ipations had only two significant sets of answers — one
relating to inflation, and the other to interest rates.
The inflation question asked for anticipated
rates of inflation. Those who answered 3 percent or
above had a mean wealth significantly greater than those
answering below 3 percent. The mean wealth of those not
anticipating any inflation was well below the other re
sponses but was not significantly different due to the
large disparity among variances. For example, those an
swering 3 percent inflation had a variance five times as
large as those answering 0 percent inflation. This
would indicate that there was a variety of wealth groups
represented in the former response and little disparity
among those who answered with the latter. The mean wealth
of those anticipating positive rates of inflation from
3-6 percent seemed to be fairly well dispersed and
hovered around $22,000. The highest mean wealth was
found in the answer 3 percent inflation ($26,451). Hence
in this analysis the mean wealth of the answers from 3-6
percent do not explain the anticipations differences
very well.
The second question dealing with anticipating
market interest rates changes found only 5 percent
120
anticipating correctly (i.e., expecting interest rates to
fall) . The mean wealth of those anticipating a decline
in interest rates was $24,932 which was not significantly
greater than those who expected them to rise ($19,136) or
those who said they expected them to remain the same
($20,538).
The failure of the "t" tests to find any signifi
cant differences among the mean wealth of the respondents
for each answer does not lend much support to the hy
pothesis that wealth and anticipations are positively
related, albeit the means do follow the hypothesized
pattern (i.e., it is largest for the correct answer and
lowest for the incorrect one) . Nevertheless, in no case
is the hypothesis in direct opposition to the results
of the tests.
The third set of significant question relates to
consumer behavior with respect to changes in the interest
rate (question 1), what to do with the new savings (ques
tion 3), and why (question 4) . In the first question,
significant differences between the mean wealth of those
responding that they did do something because of interest
rate changes was significantly greater than the mean
wealth of those saying no ($49,049 vs. $28,784). There
was also a significantly greater difference between
121
those Who said that they bought or sold various port
folio items as a result of interest rate changes and
those who said no ($62,532 vs. $28,784). This would im
ply that portfolio adjustments due to changes in interest
rates are more common for the more wealthy than for the
less wealthy. These results support the hypotheses that
with more information an individual will shift his port
folio more easily and more readily as he sees more
profitable opportunities, and also that the lower per
unit (per dollar of portfolio value), transaction costs
the easier it is (cheaper relatively) to make portfolio
adjustments.
The second question in this set has to do with
what consumers will do with additional savings. Those
indicating stocks (or trusts) and real estate were found
to have a significantly larger mean value than those who
answered bonds ($24,375 and $26,807 vs. $12,499). The
value of bonds were found to have fallen at least 11
percent in this year while the value of real estate was
estimated to grow at a rate of 5 percent. But stocks,
on the other hand, fell by almost 1-2 percent in value
during the period. The mean wealth for those responding
demand deposits, savings accounts, and the like, was
not significantly greater than that of bonds. However,
122
it was significantly lower than the mean wealth of those
who selected real estate or stocks.
This question can be treated in two ways. First,
as above, where the values of the portfolio items were
calculated for the year the decision was made (as an in
dicator of knowledge of various markets) , or it can be
treated as an anticipations question where the values are
calculated for the next year. If the latter approach is
taken, then the disastrous performance of the stock mar
ket must be taken into account. In the 1969-1970 period
the stock market lost about 14 percent in value while
bonds lost only 8 percent and real estate again rose in
value by about 6 percent. In this case, it would seem
that those preferring stocks would loose relative to
those choosing any other portfolio item. While this may
be true, only 17 percent chose the stock market while 23
percent chose real estate, and the latter did appreciate
in value significantly. The mean wealth of those se
lecting it was the highest of all choices. Furthermore,
since the stock market is probably the most volatile
portfolio item available, the average increase or de
crease in its value might be a more meaningful indicator
of its effects on the change in portfolio valuation.
The last question refers to the motives in making
the investment choice in the previous question. The "t"
123
test showed that the motivation for safety had a signif
icantly higher mean wealth value than the desire for
return ($21,221 vs. $16,085) or liquidity (21,221 vs.
$11,182). This would imply that wealthier individuals
were willing to substitute reduction in risk for return at
that time. Perhaps they foresaw the coming market de
cline, but again it is difficult to discern this from
cross-sectional data alone.
From the analysis of the contingency tables by
the mean wealth of the answers method, it seems that the
first hypothesis that wealth and information about vari
ous financial markets is supported while the second
hypothesis relating wealth and anticipation is on some
what looser ground. All answers in the first set of
questions supported the information hypothesis as well
as the responses to the questions in the behavioral set.
The second hypothesis may have run into some inexplicable
situations but the general thrust of each question can
be said to give more support to the hypothesis than it
takes away (e.g. , inflation was better anticipated by
the wealthy in general. Even though they did not pick
the correct amount of inflation, they were in the gen
eral area, in most cases. Moreover, none of the results
was in direct opposition to the hypothesis.
124
Results of the Probability
Correlation Tests
The second technique used to analyze the chi-
square tables was the probability correlation method.
This test focused upon the rows of the table while the
earlier one dealt with the columns. Both tests allowed
for the direction of the association between wealth and
particular answers to be determined. As in the earlier
analyses, only significant chi-square questions were
explored. The following are the results of the correla
tion analysis. (A summary of these results, replete
with levels at which they are significant, can be found
in Table 2.)
Question 2A of the information set (see Question
Set B) relating to interest rates on $1,000 loans was
analyzed first. Answers of 4 percent or less were nega
tively related (but only slightly) to wealth as the
correlation coefficient was small, negative and insig
nificant; r = -.031 at the .466 level of significance.
The second answer (4-6.4 percent) was more
closely related to wealth; r = -.2087. However, the cor
relation between wealth and the answer 6.6-9.5 percent
was strong, positive and significant; r = .7720 implying
the same differences in knowledge as the mean wealth
analysis (above).
125
TABLE 2
CORRELATIONS BETWEEN WEALTH AND ANSWERS (PROPORTIONAL)
TO QUESTIONS WITH SIGNIFICANT CHI-SQUARE STATISTICS
Signif
icance
sstion Answer r Levels
Information Set
Interest paid on a 1. Under 4% -.0310 .466
$1,000 loan (proposed) 2. 4-6.5% -.2087 .281
3. 6.6-9.5% .7720 .004
4. 9.6-12.5% .1791 .310
Highest rate ob 1. 3-4% . 2986 .201
tained on savings 2. 4-5% .5268 .059
3. 5-6% .9395 .001
Any noticed changes 1. Increase .6358 .024
in interest rates in 2. No change -.3988 .127
last few months?
If so what?
Anticipations Set
How large a price 1. 1%
-.1323 .358
change do you ex 2. 2% .2400 .252
pect in the next 3. 3% .6553 .020
year? 4. 5% .7751 .004
5. 6-9% .1335 .357
What do you think 1. Go up -.4448 .099
will happen to in 2. Same .1568 .333
terest rates next 3. Down .5072 .067
year?
Behavioral Set
Did you do anything 1. Yes .9201 .001
because of changes 2. No .6114 .030
in the interest rate? 3. Bought and
Sold
.9234 .001
What is wisest thing
1. Buy bonds -.6690 .017
to do with new
savings?
2. Buy real
estate
.5905 .036
3. Buy stock .7601 .005
4. D.D., T.D.,
C.D.
-8517 .001
126
TABLE 2 — Continued
Question Answer r
Signif
icance
Levels
Why that choice? 1. Safety .5402 .053
(re: above) 2. Return -.5120 .065
3. Liquidity -.6323 .025
4. Capital
gains
.1899 .300
127
Answers to the second question regarding interest
rate yields on savings showed a clear pattern relating to
wealth. Answers of 3-4 percent returns had a relatively
small (.2986) correlation coefficient, 4-5 percent a high
er correlation r = .5268, and the highest, r = .9395,
was found with the response 5-6 percent yield. Thus, as
the rate paid on savings increased so did the correlation
with wealth, indicating that the wealthy had greater ac
cess to information concerning the highest yield avail
able for savings. These findings also support the
hypothesis that wealth and information access are related
positively.
The last question in this set dealt with changes
in the rate of interest. The correct answer (increase)
was found to be positively related to wealth, r = .6358
(significant at the 2.5 percent level), while answers of
no change were negatively related to wealth (r = -.3988)
although they were not significant. A correlation with
the answer "decrease” was not performed due to the
sparsity of responses.
The second set of questions concerning anticipa
tions did not manifest very strong relationships with
wealth using the mean wealth test, albeit they were not
in opposition to the second hypothesis, i.e., the direc-
128
tion of results were correct but not significant, and in
general supported it. The correlation analysis' findings
are a bit stronger and give firmer support to the second
hypothesis.
The answers to the inflation anticipations ques
tions were negatively correlated with wealth in the
lower projected rates of inflation (r of -.1323 for 1
percent inflation, -.2400 for 2 percent). The answer,
3 percent inflation, was much more highly correlated with
wealth, r = .6553, but for an even higher rate of infla
tion, 5 percent, the correlation was even higher, .7751.
The correlation patterns suggest that higher wealth
groups tended to chose the higher (more correct) rate of
inflation. The correlations became higher and more sig
nificant as the correct amount of inflation was approach
ed.
The correct answer, 6 percent, was not found to
be strongly correlated with wealth, r = .1335, although
close but lower rates were.
The correlation analysis also shed some light on
the moot results derived from the mean wealth analysis
of the second anticipations question. The answer inter
est rates will "go up" in the next year was found to be
negatively related to wealth as the r was -.4448. "Stay
129
the same" was positive but weakly correlated, r = .1588.
"Go down" was found to have the strongest positive cor
relation with wealth, r = .5072. This pattern indicates
that wealth was related to anticipating the actual de
cline in interest rates. While the r of .5072 is not
very large, following the pattern of hypothesized sign
and relative correlation strength, it does imply a rela
tionship between wealth and good anticipations.
Both anticipations questions which supported the
second hypothesis in the earlier analysis in general
terms were found to be of much greater support using the
correlation analysis.
The correlation analysis was also useful in the
breakdown of the behavioral set of questions. Answers
to the first question of whether there had been any
changes in behavior due to changes in interest rates
when correlated with wealth rates, indicated that the
wealthy were more interest rate sensitive than the less
wealthy. The Pearson correlation coefficient for yes
answers was .9201 and .9234 for answers of bought and
sold. The "no" responses, on the other hand, showed a
relatively lower correlation with wealth, r = .6114,
than those saying they would change if interest rates
changed.
130
Answers to the second question of what to do with
additional savings also indicate an aversion to bonds and
a movement to real estate and the stock market by the
more wealthy as was found in the previous analysis. Pur
chases of bonds had a strong negative correlation with
wealth, r - -.6690, whereas investment in real estate
was strongly positive, .5905, and in the stock market,
an even stronger .7601. This seemed to fit with the
earlier analysis' findings.
A rather strong negative correlation was found
between investing extra money in demand deposits, cer
tificates of deposit and time deposits and wealth,
r = -.8517. This would suggest that those in the higher
wealth groups who anticipated inflation were trying to
stay out of fixed valued assets which lose in an inverse
manner with inflation (of any degree). These findings
suggest a better knowledge of the effects of changes in
the price level on the value of a portfolio by the
wealthier individuals.
The correlation analysis of the last question
suggests that the wealthiest individual1s prime motiva
tion for making the investment choice in the previous
question was safety, r = .5402. This was also found in
the wealth analysis (above). The correlation of wealth
131
and the response "return" as a motivation for investing
was found to be strongly negative, r = -.5120, indicating
that the wealthy were looking for safety to a much
greater extent than for returns. On the other hand, the
response "capital gains" was positively associated with
wealth, albeit only slightly, r = .1899. Liquidity was
also found to be negatively related to wealth, r = -.6323
as a motivation for investment.
These results may serve as an explanation of the
large negative correlations of wealth and demand de
posits, and so forth, in the previous question in that
the demand to gain liquidity from marginal amounts of
investment funds is much lower than that of gaining
security for the entire portfolio.
The differences in the correlations of returns
and capital gains may be due to institutional differences
as far as taxes are concerned. In the former case before
tax returns were probably considered (or taxed at ordi
nary income) and in the latter case, the opposite.
For the most part, the correlation analysis
seems to fit well with the findings, of the earlier tests
of the behavioral questions. A few other implications
about consumer behavior were also derived such as the
attitudes towards risk and reward of the wealthier
groups.
132
The analysis of significant chi-square questions
by correlating wealth and probabilities of answering in
a particular fashion, while similar to the mean wealth
analysis, provided additional insight into the tests of
the hypotheses. The sensitivity of the correlation
analysis (especially sign changes) picked up relation
ships that the "t" tests in the mean wealth analysis
could not. This was especially true in the case of the
second set of questions testing the second hypothesis,
and in light of this evidence, the support for the
second hypothesis is much stronger.
CHAPTER VI
CONCLUSIONS AND RECOMMENDATIONS
This research has been an inquiry into the wealth
distributional effects of changes in government policy to
the extent that these changes are manifest in variations
in the interest rate or the price level. Since every
individual1s portfolio valuation can be affected by
changes in these variables, they may be expected to play
a significant role in explaining wealth changes.
Theoretically, changes in the interest rate
and/or the price level should not have any redistribu
tive effects among wealth groups when market imperfec
tions do not exist. This is not the case, however, in
the real world and a redistribution among individuals in
the consumer sector may take place for one or more of
the following reasons: (1) there exist differences among
individuals' knowledge of market activities; (2) there
exist differences among individuals' ability to predict
future market outcomes; (3) there exist differences in
relative transactions costs (per dollar of portfolio
value); and (4) there may be differences in their ability
to spread risk.
133
134
Hie direction of the redistribution resulting
from the market imperfections is hypothesized to be up
ward. That is to say, changes in the interest rate or
the price level are expected to lead to wealth being re
distributed from the less wealthy to the more wealthy.
The direction of the redistribution has been determined
by exploring the four reasons mentioned above.
Two major hypotheses were developed based upon
the first two imperfections. The first is that higher
wealth and greater access to information are positively
related, and the second is that wealth and accuracy of
anticipations are also significantly and positively
associated.
In order to see how wealth is redistributed, the
individual utility maximization process is explored first
without and then with the inclusion of the four market
imperfections.
An individual's portfolio selection process is
expected to be a function not only of risk and return
but also of the amount and types of information access
he possesses. The usefulness of information access is
shown to be a declining function of time. It is demon
strated that the larger the portfolio the greater the
incentive to acquire information (i.e., more and earlier)
135
since the cost can be spread over a larger base. Optimal
purchasing times are found to be earlier and returns
greater the larger the portfolio in almost all cases.
The analysis of portfolio selection in imperfect
markets implies the existence of lags in the adjustment
of prices in the market due to lags in the dissemination
of information. Hence, a dynamic price adjustment proc
ess is suggested. Knowledge of the process whereby
information is disseminated should be transferable into
a dynamic price adjustment model. This would provide an
interesting area for further research.
The two hypotheses concerning wealth and infor
mation and anticipations are tested using questions
derived from a large consumer survey. Chi-square tests
were performed for each question to test for its associa
tion with wealth. Then each question with a significant
chi-square statistic was further investigated to see if
the mean wealth of the respondents (for each answer) was
significantly different using a student's "t" test. As
a second test, correlation of proportions of answers
made to a question and wealth levels were performed for
each answer.
The questions were broken down into three sets:
first, questions that tested for an individual's knowl
edge of present market conditions; second, for predictive
136
abilities; and, finally, for behavioral differences with
respect to changes in interest rates and the price level.
Significant chi-square statistics were found for
questions relating to interest rates and the price level
only, although questions concerning employment, business
conditions and the like were asked. This suggests that
information gaps among wealth groups are more readily
found for the financial variables than for other vari
ables more closely associated with the real sphere of
economic activity. It also suggests that information
and anticipations about and behavior toward interest
rates and the price level are associated with wealth
levels.
The mean wealth test and the correlation tests
both found that answers to each of the significant infor
mation questions (first set) were positively related to
wealth.
The second set of questions did not show a strong
relationship between wealth and various answers to the
anticipations question using the mean wealth test —
although no direct contradictions were found. Notwith
standing, the moot results eminating from the mean wealth
tests, the correlation technique did find support for a
positive association between wealth and the ability to
predict correctly.
137
Finally, both methods found initial wealth hold
ings to be a significant influence on an individual's
portfolio investment decisons and sensitivity regarding
interest rate changes. This implies variations in
amounts of market information and ability to predict mar
ket outcomes. Hence, the results of the tests seemed to
support the hypothesized positive relation of wealth
with information anticipations and economic behavior.
On the basis of the above analysis, it would seem
that initial wealth levels and changes in wealth (rela
tive) would be significantly and positively associated,
since those in the highest wealth groups would have more
information and would be able to anticipate better than
those in lower wealth levels the changes in interest
rates and the price level. Unfortunately, this hypothesis
could not be tested because wealth being defined in terms
of utility involves not only increases in portfolio
valuation but also reduction in risk (assuming individ
uals to be risk averters in general). Since risk and
reward are trade offs, it is necessary to know the pref
erence pattern (utility mapping) of risk for reward.
There is no way to determine what the level or changes
in the level of risk return are, i.e., the trade off,
using cross sectional data. (This would be an area
138
where further research would be useful.) Time series
data on surveys present difficulties because the base
keeps shifting (i.e., those in one wealth group in a
particular year might move into another the next year)
unless the same family units are re-interviewed which
is a difficult and costly undertaking. Nevertheless,
without this data patterns of trade offs between risk
and reward among and within wealth groups over time
cannot be ascertained and hence the extent of the redis
tribution also cannot be determined. The solution to
this problem would be helpful in assessing the relation
ship between initial wealth holdings and wealth redis
tribution.
Even though a good test for wealth redistribution
cannot be undertaken due to lack of data, there are many
indications that, if it goes on at all, it is in an up
ward direction with respect to changes in the interest
rate or price level. Almost all studies lead to this
conslusion. Even the researchers who found little redis
tribution (in the business sector) due to inflation admit
that it is differences in anticipations that cause wealth
redistribution. Since the accuracy of anticipations
■^R. Kessel, and A. A. Alchian, "Effects of Infla
tion," Journal of Political Economy (December 1962); and
L. DeAlessi, "The Redistribution of Wealth by Inflation:
An Empirical Test with United Kingdom Data," Southern
Economic Journal (October 1963).
139
have been shown to be positively associated with wealth,
changes in interest rates and the price level can be
expected to yield upward redistributions of wealth.
The question now to be addressed is one of de
gree. How much of a redistribution takes place as a
result of changes in government policy? This becomes a
difficult assessment without the necessary portfolio
items and their market prices. If it can be shown that
wealth is redistributed significantly by government
policies, then the following policy considerations are
in order.
First of all, it should be emphasized that what
is important for the type of redistribution of wealth
discussed here are changes in government policy that
create changes in the interest rate and price level.
Hence, whether a tight or easy money policy is pursued
or whether light or heavy taxes are levied is not of
concern. It is the changes in policy that are important.
When changes in market parameters occur, those who know
about it first and/or can anticipate the changes or the
impact of the changes will be in a position to gain most
at the expense of those who cannot anticipate as well or
who have less information.
In order to abate the wealth redistribution
140
(upward) , it would seem that the most obvious tool to
accomplish this would be stabilizing government policy
(to the extent that stable government policies would
yield stable interest rates and price levels). Since
monetary and fiscal policies attempt to achieve many eco
nomic goals besides the redistribution of wealth, at the
same time, alternatives should be considered. For ex
ample, the stabilization of monetary policy may lead to
a significant amount of new unemployment which might make
the price of wealth redistribution abatement too high.
A viable alternative might be to provide informa
tion concerning government policies and their actual and
potential effect at a very low cost (e.g., in a newspaper)
so that many more people could obtain the information
easily. This would tend to reduce the number of people
whose wealth is affected by lack of information signif
icantly. Moreover, better disclosure of government
policies and their effects might serve another useful
service at the same time, namely, to reduce uncertainty
in both the consumer and business sectors. A reduction
of uncertainty will increase the utility of the consumer.
It will also free resources previously used to determine
consumer behavior for production of goods and services.
Reducing uncertainty would also dampen the forces leading
141
to various misallocations of resources such as an excess
supply of labor for the entire economy.2 Additional re
search into the measurement of the costs of uncertainty
to society would aid the policymaker in assessing alter
native policies.
In making policy decisions, the policymaker must
weigh the importance of each alternative in terms of its
costs and benefits. This research has attempted to con
sider the effects of government policies as they are
manifest by changes in the interest rate and the price
level. An upward redistribution of wealth may result
from these policies because of an uneven distribution of
information and ability to anticipate. Including these
additional effects of government policy in this analysis
should be of assistance to the policymaker in chosing
appropriate policies as a wider range of effects are
taken into consideration.
2r . j . Barro, and H. I. Grossman, "A Generic
Disequilibrium Model of Income and Employment," American
Economic Review, Vol. 61, No. 1 (March 1971).
BIBLIOGRAPHY
142
SELECTED BIBLIOGRAPHY
Adams, P. G. "Prediction with Consumer Attitudes: The
Time Series — Cross Section Paradox," Review of
Economics and Statistics (November 1965) , 367-78.
________ . "The Size of Individual Incomes,1 ' Review of
Economics and Statistics (November 1958), 390-98.
Alchian, A. "Inflation and Distribution of Income and
Wealth," Distribution of National Income.
Edited by J. Marchal and B. Ducros. New York:
St. Martin's Press, 1968.
________ , and Kessel, R. A. "Redistribution of Wealth
Through Inflation," Science (September 4, 1959),
535-39.
Almond, Al Somarrie, and Miller, H. P. "State Differ
entials in Income Concentration," American
Economic Review (March 1967), 59-72.
Anderson, Theodore W. An Introduction to Multivariate
Statistical Analysis. New York: John Wiley &
Sons, Inc., 1958.
Bach, G. L., and Ando, A. "The Redistribution Effects
of Inflation," Review of Economics and Statis
tics (February 1957), 1-13.
Barlow, Robin; Brazer, Harvey E.y and Morgan, James N.
Economic Behavior of the Affluent. Washington,
D.C.: The Brookings Institution, 1966.
Barro, R. J., and Grossman, H. I. "A Generic Disequi
librium Model of Income and Employment,"
American Economic Review. Vol. 61, No. 1
(March 1971), 82-93.
Brimmer, Andrew P. "Inflation and Distribution in the
U.S.," Review of Economics and Statistics
(February 1971), 37-48.
Bronfenbrenner, M. Income Distribution Theory.
Illinois: Aldine Publishing Company, 1971.
143
144
Bronfenbrenner, M., and Holzman, F. D. "A Survey of In-
flation Theory," American Economic Review
(September 1963), 593-661.
Brownlee, 0., and Conrad, A. "Effects Upon the Distri
bution of Income of a Tight Money Policy,"
Commission on Money and Credit. Stabilization
Policies. Englewood Cliffs, N.J.: Prentice-
Hall, Inc., 1963.
Budd, Edward C. "Postwar Changes in the Size Distribu
tion of Income in the U.S.," American Economic
Review (May 1970), 247-60.
________, and Seiders, David. "The Impact of Inflation
on the Distribution of Income and Wealth,"
American Economic Review (May 1971), 128-38.
Bureau of the Budget. Family Income Distribution Sta
tistics Published by Federal Agencies. Statis
tical Evaluation Report No. 5 (December 1964).
Bureau of Labor Statistics. Study of Consumer Expendi
ture. Incomes, and Savings. 1950. Edited by
Irwin Friend. Philadelphia: University of
Pennsylvania Press, 1959.
________. Survey of Consumer Expenditure. Reports
273-293 (May 1966) .
Burger, Albert E. "Effects of Inflation (1960-1968),"
Federal Reserve Bank of St. Louis (November
1969), 25-36.
Claycamp, Henry. "The Composition of Consumer Port
folios." Unpublished Ph.D. dissertation,
University of Illinois, 1961.
Clower, Robert A. "A Reconsideration of the Micro
foundations of Monetary Theory," Western
Economic Journal (December 1968), 1-8.
Conard, J. W. "The Causes and Consequences of Infla
tion," Commission on Money and Credit. Infla
tion Growth and ESnployment. Englewood Cliffs,
N.J.: Prentice-Hall, Inc., 1964.
Conover, W. J. Practical Nonparametric Statistics.
New Yorfc: John Wiley & Sons, Inc., 1971.
145
Crockett, Jean, and Friend, Irwin. “Consumer Investment
Behavior," Determinants of Investment Behavior.
N.B.E.R. Study. Edited by R. Ferber. New York:
Columbia University Press, 1967.
Datta, U. “The Distribution of Income from Wealth,“
Papers on National Income and Allied Topics.
Vol. III. Indian Association for Research in
National Income and Wealth. Edited by N. S. R.
Sastry, et al. New York: Asia Publishing House,
1965.
Davies, J. C. 1 1 Some Relations Between Events and Atti
tudes," American Political Science Review
(September 1952), 777-89.
DeAlessi, L. “The Redistribution of Wealth by Inflation:
An Empirical Test with United Kingdom Data,“
Southern Economic Journal (October 1963), 113-27.
DeWitt, L. B. “Inflation and the Distribution of Unem
ployment by Income Class." Unpublished Ph.D.
dissertation, Syracuse University, 1970.
Federal Reserve Board. “Reports on the Survey of Con
sumer Finances,“ Federal Reserve Bulletins,
1946-1959. ‘
Ferrer, A. "Income Distribution: Panel Discussion,"
Commission on Money and Credit: Inflation.
Growth, and Employment. Englewood Cliffs, N.J.:
Prentice-Hall, Inc., 1964.
Foster, Edward. "Costs and Benefits from Inflation,"
Studies in Monetary Economics. Federal Reserve
Bank of Monneapolis (March 1972).
Francis, D. R. "Monetary Policy and Inflation," Federal
Reserve Bank of St. Louis (June 1969), 14-23.
Goldsmith, Raymond, et al. A Study of Savings in the
United States. Vol. III. Princeton: Princeton
University Press, 1956.
________ , and Lipsey, Robert E. Studies in the National
Balance Sheet of the United States. N.B.E.R.
Study. Princeton: Princeton University Press,
1963.
146
Goldsmith, S. F. "Low Income Families and Measures of
Income Inequality," Review of Social Economy
(March 1962), 1-25.
________. "Statistical Information on the Distribution
of Income by Size in the United States,"
American Economic Review (May 1950), 321-41,
Gordon, Donald F., and Hynes, Allan. "On the Theory of
Price Dynamics," Microeconomic Foundations of
Employment and Inflation Theory. Edited by
Edmund Phelps, et al. New York: W. w. Norton
& Company, 1970.
Harris, Seymour. "The Incidence of Inflation" or "Who
Gets Hurt?" Study Paper No. 7. Study of Em
ployment. Growth and Price Levels. Joint Eco
nomic Committee, 86th Cong., 1st sess. Washing
ton, D.C.: Government Printing Office, 1959.
Hollister, Robinson, and Palmer, John. "The Impact of
Inflation on the Poor." Discussion Paper No.
40-69, Institute of Research and Poverty,
University of Wisconsin, 1969.
Internal Revenue Service. Statistics of Income: Indi
vidual Income Tax Returns. Washington, D.C.:
Government Printing Office, 1967-1968.
Juster, Thomas. Anticipations and Purchases: An Analysis
of Consumer Behavior. N.B.E.R. Study. Prince
ton: Princeton University Press, 1964.
Katona, George. "Attitudes Toward Fiscal and Monetary
Policy," Public Policy (Winter 1970), 420-441.
"Changes in Consumer Expectations and Their
Origin," The Quality and Significance of Antic
ipation Data. N.B.E.R. Study. Princeton:
Princeton University Press, 1960.
________. "Expectations and Decisions in Economic Be
havior," The Policy Sciences. Edited by D.
Lerner and H. Lasswell. Stanford: Stanford
University Press, 1951.
Psychological Analysis of Economic Behavior.
New York: McGraw-Hill Book Company, 1951.
147
_. The Powerful Consumer. New York: McGraw-
Hill Book Company, 1960.
_. "The Predictive Value of Data on Consumer
Attitudes," Consumer Behavior. Vol. II. Edited
by Lincoln Clark. New York: New York University
Press, 1955.
_________, et al. Survey of Consumer Finances (1960,
1963, 1966-1970). Ann Arbor: University of
Michigan Press.
_________, and Lansing, J. B. "The Wealth of the
Wealthy," Review of Economics and Statistics
(February 1964), 1-13.
_________, and Mueller, E. Consumer Attitudes and Demand
1950-1952. Ann Arbor: University of Michigan,
1951.
________ , and Mueller, Eva. Consumer Expectations 1953-
56. Ann Arbor: University of Michigan Press,
1956.
. Consumer Response to Income Increases. Wash
ington, D.C.: Brookings Institution, 1960.
Kendall, Maurice 0. The Advanced Theory of Statistics.
Vol. I. London: C. Griffin & Company, 1948.
________ , and Stuart, A. The Advanced Theory of Sta
tistics. Vols. II and III. New York: Hafner
Publishing Company, 1961, 1966.
Kessel, R. "Inflation-Caused Wealth Redistribution: A
Test of a Hypothesis," American Economic Review
(March 1956), 128-41.
_________, and Alchian, A. A. "Effects of Inflation,"
Journal of Political Economy (December 1962),
521-37.
_________, and . "The Meaning and Validity of
the Inflation Induced Lag of Wages Behind
Prices," American Economic Review. Vol. L
(March 1960), 43-66.
148
Klein, Lawrence. "Assets, Debts, and Economic Behavior,"
Studies in Income and Wealth. Vol. 14. N.B.E.R.
Study. New York, 1951.
________ , ed. Contributions of Survey Methods to Econom
ics. New York: Columbia University Press, 1954.
Lansing, John, and withey, S. "Consumer Anticipations:
Their Use in Forecasting Consumer Behavior,"
Studies in Income and Wealth. Vol. 17. N.B.E.R.
Study. Princeton: Princeton University Press,
1955.
Lerner, Eugene M., and Carleton, Williard T. A Theory of
Financial Analysis. New York: Harcourt, Brace
& World, 1966.
Litner, John. "Security Prices, Risk and Maximum Gains
from Diversification," Journal of Finance
(December 1965), 587-615. — -
Lydall, Harold, and Lansing, John. "Distribution of Per
sonal Income and Wealth,1 1 American Economic
Review (March 1959), 43-67.
Markowitz, Harry M. Portfolio Selection: Efficient
Diversification of Investments. New York:
John Wiley & Sons, Inc., 1959.
Maxwell, A. E. Analyzing Qualitative Data. New York:
John Wiley & Sons, Inc., 1961.
Miller, Robert F., and Watts, Harold W. "A Model of
Household Investment in Financial Assets,"
Determinants of Investment Behavior. N.B.E.R.
Study. Edited by Robert Ferber. New York:
Columbia University Press, 1967.
Moore, Basil T. An Introduction to the Theory of
Finance. New York: Free Press, 1968.
Morgan, J. "The Anatomy of Income Distribution,"
Review of Economics and Statistics (August 1962),
270-83.
Mueller, Eva. "Consumer Attitudes: Their Influence and
Forecasting Value," The Quality and Significance
of Anticipation Data. N.B.E.R. Study. Prince
ton: Princeton University Press, 1964.
149
"Consumers' Attitudes Toward Saving and Their
Investment Preferences," The Analyst1s Journal
(November 1957), 3-9.
"Consumer Reactions to Inflation," Quarterly
Journal of Economics (May 1959), 246-62.
"Effects of Consumer Attitudes on Purchases,"
American Economic Review (December 1957), 946-65.
"Ten Years of Consumer Attitude Surveys: Their
Forecasting Record," Journal of the American Sta
tistical Society (December 1963), 899-918.
"The Impact of Unemployment on Consumer Con
fidence, " Public Opinion Quarterly (Spring 1966),
19-32.
Perry, J. Unemployment. Money Wage Rate, and Inflation.
Cambridge, Mass.: MIT Press, 1966.
Pesek, B. p. "A Comparison of the Distributional Effects
of Inflation and Taxation," American Economic
Review (March 1960), 147-153.
_. "Macroeconomic Theories of Income Distribu
tion: Discussion," American Economic Review
(May 1961), 88-91.
Prichard, W. C. "Inflation, Expectations and Wealth
Distribution." Unpublished Ph.D. dissertation,
Michigan State University, 1971.
Projector, Dorothy. "Consumer Asset Preference,"
American Economic Review (May 1965), 227-51.
_. "Re-examination of Saving Theory in Light of
FRB Surveys of Financial Characteristics of Con
sumers," American Statistical Association,
Business (1968), 56-60.
Saving, T. R. "Portfolio Choice and Monetary Policy,"
Journal of Money and Credit (May 1970), 258-67.
Scheffe, Henry. The Analysis of Variance. New York:
John Wiley & Sons, Inc., 1959.
Scherf, H. "Inflation and Income Distribution,"
Weltwirtschaft Archiv (June 1968), 258-70.
150
Schultz, Charles L. "Recent Inflation in the U.S.,"
Study of Employment, Growth and Price Levels.
Study Paper No. 1. Joint Economic Committee,
86th Cong., 1st sess. Washington, D.C.:
Government Printing Office, 1959.
Schultz, R. R. "On the Measurement of Income Inequality,"
American Economic Review (March 1951)r 107-22.
Schultz, T. P. "A Study of the Statistics on the Size
Distribution of Personal Income in the U.S.,"
The Distribution of Personal Income. Joint
Economic Committee, U.S. 88th Cong., 2d sess.
Washington, D. C.: Government Printing Office,
1964.
________ . "Secular Trends and Cyclical Behavior of In
come Distribution in the U.S., 1944-1965," Six
Papers on the Size Distribution of Income and
Wealth. N.B.E.R. Study. Edited by Lee Soltow.
New York: Columbia University Press, 1969.
Scitovsky, Tibor, and Scitovsky, Anne. "Inflation Versus
Unemployment: An Examination of Their Effects,"
Commission on Money and Credit, Inflation, Growth
and Employment. Englewood Cliffs, N.J.:
Prentice Hall, Inc., 1963.
Smith, James, and Morgan, James N. "Variability of Eco
nomic Well Being and Its Determinants," American
Economic Review (May 1970), 286-95.
________ , and Calvert, S. K. "Estimating the Wealth of
Top Wealth Holders from Estate Tax Returns,"
Proceedings of Business and Economic Statistics.
Washington, D.C.: American Statistical Associa
tion, 1965.
Soltow, L. "The Share of Lower Income Groups in In
come, " Review of Economics and Statistics
(November 1965), 429-33^
Summers, Robert. An Econometric Investigation of the
Lifetime Size Distribution of Average Annual
Income. Discussion Paper No. 9. New Haven:
Cowles Foundation, 1956.
Swan, N. M. "Inflation and the Distribution of Income."
Unpublished Ph.D. dissertation, University of
Pennsylvania, 1970.
151
Tait, A. A. "A Simple Test of the Redistributive Nature
of Price Changes for Wealth Owners in the U.S.
and U.K." Review of Economics and Statistics
(November 1967), 651-54.
Thurow, Lester. "Analyzing the American Income Distri
bution, " American Economic Review (May 1970),
261-69.
Tintner, Gerhard. Econometrics. New York: John Wiley
& Sons, Inc., 1952.
Tobin, James. "Liquidity Preference as Behavior Toward
Risk," Review of Economic Studies (February 1958)
65-86.
________ . "On the Predictive Value of Consumer Inten
tions and Attitudes," Review of Economics and
Statistics (February 1959), 1-11.
________ . "The Theory of Portfolio Selection," The
Theory of Interest Rates. Edited by Mann and
Brechling. London: Macmillan & Co., Ltd., 1965.
U.S. Department of Commerce. Survey of Current Business.
Vols. 46 through 50. Washington, D.C.: Govern
ment Printing Office, 1966-1970.
Watts, Harold W. "Long Run Income Expectations and Con
sumer Saving," Studies in Household Economic
Behavior. Vol. I. New Haven: Yale University
Press, 1958.
_______ , and Tobin, James. "Consumer Expenditures and
the Capital Account," Studies of Portfolio Be
havior . Edited by Hester Tobin, and James Tobin.
New York: John Wiley & Sons, Inc. , 1967.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A Dynamic Model For Predicting The Effects Of Fiscal Policy Measures

PDF

The Effect Of Wealth On Consumption: Empirical Tests For Significance Andlags

PDF

Planning For Stability And Self-Reliance: An Evaluation Of Policy Approaches For India

PDF

The Interaction Among Financial Intermediaries In The Money And Capital Markets: A Theoretical And Empirical Study

PDF

An Econometric Analysis Of The Influence Of Money Supply And And Money Demand Relations In The Determination Of National Income

PDF

Valuation Adjustment In Input-Output Tables, Economic Structure, And Economic Systems: A Comparative Empirical Analysis

PDF

The Economic Impact Of Savings And Loan Associations On Residential Construction

PDF

Libya's Petroleum Policy

PDF

European Economic Integration And African States

PDF

A Constrained Price Discriminator: An Application To The U.S. Post Office

PDF

The Role Of The Manager Of The Federal Reserve System Open Market Account: 1951-1961

PDF

A Simple Macroeconomic Model For The Chilean Economy

PDF

On The Theory Of Value And Market Syndicalism

PDF

A Theoretical Analysis Of The Dynamic Aspects Of Economic Integration Of Less Developed Countries

PDF

Export Instability And Economic Development: A Statistical Verification

PDF

Interstate Natural Gas Supply And Intrastate Market Behavior

PDF

The Anomaly Of External Effects: A History Of Concept Development

PDF

An Econometric Overhead Model Of An Aerospace Firm

PDF

Cross-Sectional Analysis Of Manufacturing Production Functions In A Partitioned 'Urban-Rural' Sample Space

PDF

An Estimate Of Some Basic Parameters Of The Moroccan Economy

Asset Metadata

Creator Ponchick, Elliot Allan (author)

Core Title Some Effects Of Monetary And Fiscal Policy On The Distribution Of Wealth

Degree Doctor of Philosophy

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag economics, general,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor DePrano, Michael E. (committee chair), Beling, Willard A. (committee member), Tintner, Gerhard (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-809733

Unique identifier UC11364256

Identifier 7318836.pdf (filename),usctheses-c18-809733 (legacy record id)

Legacy Identifier 7318836

Dmrecord 809733

Document Type Dissertation

Rights Ponchick, Elliot Allan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA